Text2026
Ricursive Intelligence · USA
A dense (not MoE) model with 65B parameters, pretrained on 12 trillion tokens, with a context of up to 1 million tokens. It is built for programming and agents: by the authors' own test it scores 80.52 on Terminal-Bench 2.1 and beats much larger American open models (Reflection Beam 501B with 80.1, Meta Muse Spark Open 816B with 76.4, NVIDIA Nemotron 3 Ultra 550B with 56.4). The authors started from the Olmo 3 recipe and claim a 16 to 55 times gain in training efficiency on code. It was designed for NVIDIA graphics cards. BF16 and NVFP4 are published for the instruction-tuned version, plus a BF16 base. This is a preview release, and the figures are the authors' own, not independent.
- an agent for programming in the terminal and a repository
- long tasks with a context of hundreds of thousands of tokens
- your own tuning on top of an open base model
- Sizes
- 65B (dense)
- Hardware
- from: Cluster
TextGGUF2025–2026
Moonshot AI · China
Very large Moonshot MoE models for agentic work. K3 (2.8 trillion parameters) was the largest open model at release, with up to 1M tokens of context and image understanding; K2.7-Code is built for programming. On November 4 the weights of K3.1 Pro were opened (2.8 trillion parameters, 104B active, 1M tokens of context, images): a refinement of K3 for programming, 3D and financial research; by the tests of Moonshot itself, Terminal-Bench 4.0 is 65.7%. The weights are in MXFP4, about 1.56 TB, run through vLLM and SGLang, and need a serious cluster.
- Multi-step agents: search, data collection, reports
- In-depth document analysis
- Help for developers
- Sizes
- 16B-A3B – 2.8T-A104B
- Hardware
- from: 1 GPU
Image + text2026
Microsoft · USA
A compact 4B model that operates a computer by itself: it looks at a screenshot and clicks, and when that is more efficient it switches to the command line and edits files with code. Both roles are in one model, with no external orchestrator. On the OSWorld-Verified test (50 steps) it scores 66.0%, which the authors say is the best result among 4B models, and 87.5% on operating-system tasks. It was trained from the intermediate Mage-VL in five stages, with a set of almost 294 thousand synthetic tasks over 6,915 skills. The context for training and running is 32,768 tokens, and the architecture holds up to 262,144. It suits office automation and assistants on your own computer. The figures come from the authors.
- an agent that works in applications through screenshots and the command line
- automating office tasks on your own computer without a cloud
- a prototype assistant that operates the desktop
- Sizes
- 4B
- Hardware
- from: Laptop
TextOllama2024–2026
NVIDIA · USA
NVIDIA models for agents and reasoning, optimized to run fast on its GPUs. Nemotron 3 is a Mamba and MoE hybrid from 4B to 550B; Nano Omni handles video, audio and images (English only). The new Nemotron 3.5 Super VL (120B, 12B active, up to 1M-token context) reads text, images and video up to two minutes, with a focus on documents and agents; the compressed Bolt (74B, 9B active) fits into 24 GB of video memory. Super VL languages: English, Spanish, French, German, Italian, Japanese. Dory 1.0 (4B, up to 256K-token context, English) runs the same block of layers several times, and the number of passes can be changed per task: more passes give more accuracy without a new model and without long reasoning in the text.
- Agents with tool calling
- Reasoning and calculation tasks
- Answers based on long documents
- Sizes
- 4B – 550B-A55B
- Hardware
- from: Laptop
Text2026
Nanbeige · China
A 33B model that the authors call the senior teacher for future lightweight agents: they plan to train small models (for example 3B) on it. It is strong in programming and office tasks: on the JobBench test across 35 professions it beats the baseline in most, according to the authors. Context up to 512 thousand tokens, reasoning can be switched on and off, and tool calling works best in the XML format. The authors compare it with Qwen3.8-27B, Muse-Glimmer-30B and Gemma 4 31B and say it leads on most tests; the figures are their own. A base version is published too.
- an agent for programming and working with tools
- office tasks by profession: reports, tables, document analysis
- a base model for your own fine-tuning
- Sizes
- 33B
- Hardware
- from: 1 GPU
Text2025–2026
Ant Group (inclusionAI) · China
An Ant Group family: Ling for standard models, Ring for reasoning ones. There are trillion-parameter flagships and the efficient Ling-3.0-tiny, which needs only 1.3 billion active parameters. The new Ling-3.1-flash (560B, 25B active, 1M-token context) targets agents, office tasks and development. The improved Ling-3.1-flash-1019 (MIT) raised the Artificial Analysis index from 41 to 44 by the authors estimate.
- Corporate assistant
- Agents for office processes
- Financial analytics (Fin version available)
- Sizes
- 7.9B-A1.3B – 1T
- Hardware
- from: Laptop
Text2026
XHToken (the Spark project) · not stated
A compact 4B model to run right on a device, focused on agents: tool calling, multi-step planning, programming, long context. Version 2.6 came about a month after 2.5 with the same architecture and improved only through further training. The reasoning budget is adjustable: none, low, high. The authors state that it holds its own against models of around 9B. Languages on the card: English and Chinese, Russian is not stated. Running it needs the model's own code in transformers. The figures come from the authors. At the end of October a 1.7B size was added (INT4, INT8, FP8).
- an agent with tool calling on a home computer or laptop
- a local programmer assistant that does not need a big server
- AI built into an app with a choice of how much the model thinks
- Sizes
- 1.7B и 4B
- Hardware
- from: Laptop
Computer-use agentsGGUF2025–2026
Ant Group (inclusionAI) · China
An Ant Group family for finding elements on screen and completing tasks in phone and computer interfaces. UI-Venus-2 was specifically trained to refuse dangerous actions. In October LBS-27B came out (Apache 2.0, based on Qwen3.8-27B): it learns to pick and zoom into the needed parts of an image on interface markup, and carries this skill over to ordinary questions about pictures.
- Automating actions in mobile apps
- Filling in forms in web interfaces
- UI autotests
- Sizes
- 2B – 72B
- Hardware
- from: Laptop
Text analysisRU2026
vLLM Semantic Router project · international project
An open family of models for an AI gateway: they answer a request with probabilities without generating text. They detect topic and urgency, jailbreak attempts, harmful requests, personal data and whether a tool is needed. Small versions run on a CPU, Vela understands 17 languages including Russian. The new d3 (27B) also reads several images and videos, and d3-turbo is, by the authors' own estimate, 4 points more accurate on text. The newest in the line, d3-ultra, came out on October 24.
- routing requests between models and teams
- checking a request for a jailbreak attempt or harmful content before the main model
- finding personal data with positions in a text
- Sizes
- 0.3B – 27B
- Hardware
- from: Laptop
TextRUGGUF2025–2026
MiniMax · China
Large MoE models with very long context (up to 1M tokens for Text-01 and M3). M3 is multimodal and understands images. In October the open weights of M3.1-Flash came out (about 428B, 23B active, up to 1M tokens of context, text, images and video): a model for programming, agents and computer use, sped up by sparse attention. MiniMax Community License: free for non-commercial use; commercial use needs the caption "Built with MiniMax M3.1-Flash" and a one-time notice to MiniMax, and a separate permission when the product earns $20 million a year or more; military use is prohibited. The benchmark figures come from MiniMax itself. Licenses differ greatly from version to version.
- Analysis of large document archives in a single request
- Agents with tools
- Help for developers
- Sizes
- 230B-A10B – 456B-A46B
- Hardware
- from: Cluster
Search and RAGRU2025–2026
NVIDIA · USA
NVIDIA embeddings for search and RAG. Nemotron-3-Embed, released in 2026, is under the permissive OpenMDW license and works in many languages. In October Nemotron-3.5-Embed (1B and 8B, OpenMDW 1.1) came out: it finds documents by text and by the image of a page with tables, charts and infographics; for text it was tested in 34 languages, Russian among them. The 8B is built on Ministral 3.
- Search across corporate documents
- RAG for chatbots and assistants
- Search across images and pages (VL versions)
- Sizes
- 1B – 8B
- Hardware
- from: Laptop
Text2026
KRAFTON AI · South Korea
A model from the Korean KRAFTON (the publisher of PUBG) with 30B parameters, 4B active, trained from scratch. It is strong in Korean conversation and agentic tasks: it can reason and call tools, and works with Hermes Agent, Pi and OpenCode. Training languages: Korean, English, Japanese, Chinese, Spanish, French, Thai, and Russian is not among them. There are compressed MXFP8 and NVFP4 versions and an auxiliary model for faster generation. This is a trial release (preview), and the specifications may change before the final one. The weights are issued on request on Hugging Face, and the benchmark figures are the authors' own.
- a Korean-language assistant and customer support
- an agent with tool calling on your own server
- game and media projects for a Korean audience
- Sizes
- 30B-A4B
- Hardware
- from: 1 GPU
Code2026
Recursive · not stated
An open expert model (39.8B parameters, 2.5B active per token) for coding agents: it digs into someone else's repository, fixes bugs, writes functions and tests, and can reason and call tools over several steps. Trained from scratch, 262 thousand token context, text only. It weighs about 80 GB, and the authors run it on a single B300 or H200 card through SGLang. By the authors' own tests: Terminal-Bench 2.1 54.9%, SWE-bench Pro v2 89.7%, SWE-bench Verified 74.2%. Part of the comparisons come from other developers' own reports, and the test conditions differ between sources.
- an agent for fixing bugs and extending a repository in the terminal
- writing tests and new functions from a description
- a local developer assistant on your own server
- Sizes
- 39.8B-A2.5B
- Hardware
- from: 1 GPU
Text analysis2026
Contrastive-LM (per the announcement: Stanford together with NVIDIA) · USA
An open fast-choice model for agents: it takes a description of a situation and a set of options and returns the probability of each without generating text. It can answer yes or no, pick from a list and give a score on a scale. Actions and state are encoded separately, so action vectors can be computed in advance: by the authors' measurements, with a choice among about a thousand options it is up to 13 times faster than the Jev analogue. After fine-tuning it works as a verifier that picks the best of several solutions of a coding agent (DeepSWE 81.6%, Terminal-Bench 2.1 87.6% by their measurements). It needs Linux, an NVIDIA card and Qwen3-8B in vLLM, and the card language is English.
- fast choice of an agent's next action from given options
- checking and ranking several solutions of a coding agent
- routing requests: department, urgency, a score on a scale
- Sizes
- 8B (a frozen Qwen3-8B plus two small projection heads)
- Hardware
- from: 1 GPU
CodeOllama2026
DeepReinforce · not disclosed
Models for agentic development: they build their own plan and scaffolding for a task and execute it in the terminal. Fine-tuned from Qwen 3.5 and Gemma 4; work with Claude Code, OpenHands and similar tools. In October Ornith-1.5-122B (MoE, MIT) came out: according to the authors it is stronger and faster than the previous ones, writes shorter and is meant for a single DGX-class card; compressed FP8, NVFP4, GGUF and MLX versions exist. The figures come from the authors.
- A developer agent in the terminal
- Fixing bugs from a task description
- Understanding and extending a large repository
- Sizes
- 9B – 397B
- Hardware
- from: 1 GPU
Text analysis2026
Tencent Hunyuan · China
A model that runs the long-term memory of an AI agent: it decides what from a chat, work tasks or code is worth remembering, in what form, and how new information changes what is already stored. It replaces the general-purpose model inside the memory systems Mem0, MemOS and Tencent's own Hy-Memory. Languages on the card: English and Chinese. The benchmark figures are Tencent's own.
- an assistant's memory about a customer between conversations
- remembering the course of work on a project and code
- updating stored facts when new data arrives instead of duplicates
- Sizes
- 30B-A3B
- Hardware
- from: 1 GPU
TextOllama2026
ConwayResearch (Underdog) · not stated
Qwen3.8-27B squeezed to 2 bits into a single 7.9 GB GGUF file instead of 54 GB, with the focus on keeping tool calling intact. By the authors' measurements (their Underdog Bench), tool calling is 88 against 84 for the full model, parallel calls 42 against 35, and on average 96% of the full model's quality across nine tests. It runs on plain llama.cpp and Ollama, thinking mode can be switched on and off, and it understands images with a separate vision file (about 0.6 to 0.9 GB). Language stated on the card: English.
- an agent with tool calling on a home graphics card
- a local coding assistant with functions
- analysing pictures with the vision file attached
- Sizes
- 27B (7.9 GB file)
- Hardware
- from: Laptop
Text2026
Reflection AI · USA
An open model from the American company Reflection with 501B parameters (23B active, a 1 million token context), text only, mainly English. Stated as economical in reasoning: according to the authors it needs about three times fewer tokens for thinking at comparable quality. Strong in programming, tool calling and science. Weights in BF16, FP8 and NVFP4. The benchmark figures are the authors' own.
- an agent for solving tasks in a repository and a terminal
- long tasks with tools and search (via MCP)
- solving scientific and mathematical problems with reasoning
- Sizes
- 501B-A23B
- Hardware
- from: Cluster
TextOllama2023–2026
DeepSeek · China
DeepSeek's flagship line: from the first 7B/67B to V4-Pro with 1.6 trillion parameters. Closed-model quality under an open MIT license; V4-Flash-Vision-Exp and V4.1-Flash understand images, context up to 1M tokens. The fresh V4.1-Flash-1017 answers shorter and faster than the previous Flash: by the measurements of DeepSeek itself, 77.0 vs 74.2 on DeepSWE v1.1 and 42.3 vs 31.2 on Terminal-Bench 4.0.
- Employee assistant on your own server
- Analysis of long contracts and reports
- Agents that work with tools and APIs
- Sizes
- 7B – 1.6T-A49B
- Hardware
- from: Laptop
Text2026
Meta Superintelligence Labs · USA
Meta's largest open model: 816B parameters, 42B active per token, a 1 million token context. It understands text, images, video and audio and answers in text. Built for programming and long agentic tasks. Trained on data in more than 100 languages. Access to the weights on Hugging Face is granted manually.
- an agent for working with code and long tasks with tool calling
- analyzing screenshots, charts, documents, video and recordings
- generating training data for other models
- Sizes
- 816B-A42B
- Hardware
- from: Cluster
RerankersRU2025–2026
NVIDIA · USA
A small 1B reranker from NVIDIA. The vl version also takes document pages as images, not just text. The new 8B (Nemotron-3.5-Rerank) was evaluated on 34 languages, including Russian, and accepts text, page images or both, including charts and tables.
- Reordering passages before an AI assistant answers
- Sorting retrieved scan and PDF pages
- Search across internal policies and instructions
- Sizes
- 1B and 8B
- Hardware
- from: Laptop
Text2026
Shanghai Artificial Intelligence Laboratory · China
An agent model for long tasks: searching and gathering evidence, research, development, preparing reports and presentations, analyzing vulnerabilities in an authorized environment. A preview based on GLM-5.2, 256K-token context, text only. An FP8 version exists.
- deep search and research on a topic with a verifiable result
- developing programs and preparing reports and presentations
- checking and fixing vulnerabilities in an authorized environment
- Sizes
- 744B (MoE, based on GLM-5.2)
- Hardware
- from: Cluster
Code2026
Microsoft · USA
A compact agent for working with repository code, based on Qwen3.5-4B: searches the project, edits files, runs tests and prepares a patch. Trained only with reinforcement learning on about 1500 synthetic tasks. Works through the light Leaf harness with five tools.
- fixing bugs in a repository from a task description
- navigating an unfamiliar codebase
- preparing patches with test runs
- Sizes
- 4B
- Hardware
- from: Laptop
Computer-use agents2026
Tencent · China
An open agent for operating a desktop: looks at live screenshots, reasons and outputs clicks and keyboard input. You can show a workflow once and the agent carries it over to a new task. An external environment is needed to execute the actions.
- automating work in programs with no API
- repeating a workflow from one shown example
- testing interfaces
- Sizes
- 27B
- Hardware
- from: 1 GPU
Image + text2026
Interfaze · not stated
One model for developers: reads documents and tables, transcribes speech with speaker labels, finds objects in an image and answers strictly by a JSON schema you supply. Inside is a reasoning core and a set of narrow specialist models.
- parsing documents and tables with text coordinates
- transcribing recordings with speaker labels
- extracting data from documents into a given JSON schema
- Sizes
- a model that needs 80 GB of video memory (the core is based on Qwen3.8-27B)
- Hardware
- from: 1 GPU
Computer-use agents2026
Beijing Academy of Artificial Intelligence (BAAI) · China
An agent model of 27B (based on Qwen3.8-27B) for long tasks: it proposes a solution, measures the result, analyzes errors and revises the solution over several rounds. A 262K-token context. Stated as a research model.
- iterative improvement of code and algorithms
- machine learning tasks with a verifiable result
- deep search and gathering information on a topic
- Sizes
- 27B
- Hardware
- from: 1 GPU
Text analysis2026
AutoTrust AI · Singapore
Open decision models: for yes/no, pick-one and rate-0-to-5 questions they answer with probabilities in one pass, and can also reason when needed. Not affiliated with the closed Jev from TypeSafe.
- routing requests to the right team
- checking an agent action before it runs
- rating urgency and tone of messages
- Sizes
- 9B – 35B
- Hardware
- from: 1 GPU
Text2026
Aleph Alpha · Germany
A European open reasoning model for German and English. A 78B mixture-of-experts with 3.5B active per token, a 262K-token context (tested up to a million) and tool calling.
- answers over long documents and RAG
- an agent that calls tools
- extracting structured data from documents
- Sizes
- 78B-A3.5B
- Hardware
- from: Cluster
Text analysisOllama2026
Cloudflare · USA
Decision models: they do not write text but read a situation (text, JSON, an image, video frames, and for Clef-Omni also audio) and return in one pass the probability of each answer to typed questions: yes or no, pick one, rate on a scale. The agent gets ready numbers instead of parsing free text.
- sorting requests and tickets into categories with a probability
- checking a document or photo against a list of questions (is everything stated, is there a violation)
- choosing the next step of an agent without parsing a text answer
- Sizes
- 27B, 9B (Clef-flash) and 30B-A3B (Clef-Omni)
- Hardware
- from: Laptop
Search and RAGOllama2025–2026
Google · USA
Small multilingual embeddings from Google that run on a phone or laptop without internet. The second version also understands images, video and audio in one vector space.
- On-device document search
- RAG without sending data outside
- Search over photos, video and audio by a text query
- Sizes
- 300M и 740M (EmbeddingGemma 2)
- Hardware
- from: Laptop
TextGGUF2025–2026
StepFun · China
StepFun MoE models built for fast, low-cost work: with 196 billion parameters, Step-3.5/3.7-Flash use about 11 billion per token. Compact Step3-VL-10B for images and voice Step-Audio 2 mini are available. The flagship Step 5 Preview (600B, 27B active, up to 1M-token context, text and images as input) is built for agents, programming and finance tasks.
- High-load agents
- Analysis of documents with diagrams and screenshots
- Help for developers
- Sizes
- 8B to 600B
- Hardware
- from: 1 GPU
Text analysis2024–2026
Urchade Zaratiana and Fastino AI · France / USA
Finds the entities you need in text without training: just list what to look for (name, amount, date). GLiNER2 also classifies text. Multilingual versions understand Russian.
- Extracting names, amounts and dates from emails and contracts
- Parsing requests into CRM fields
- Classifying requests by topic
- Sizes
- about 50M to 500M
- Hardware
- from: Laptop
TextOllama2023–2026
Shanghai AI Laboratory · China
Models from Shanghai AI Laboratory. The early InternLM line is general-purpose; the new Intern-S1/S2 is scientific: it understands formulas, molecules, charts and images.
- Research assistant: papers, formulas, data
- Analysis of scientific and technical documents
- Corporate chat on small models
- Sizes
- 1.8B – about 1T
- Hardware
- from: Laptop
Text2024–2026
MBZUAI, Institute of Foundation Models (IFM, LLM360 project) · UAE
Fully open models from the UAE: data, training code and intermediate checkpoints are published along with the weights. K2-Horizon (2026) spans 0.9B to 375B with context up to 512K tokens.
- Reasoning, maths and technical questions
- Analysing long documents
- Agents and writing code
- Sizes
- 0.9B – 375B-A23B
- Hardware
- from: Laptop
TextGGUF2025–2026
Renmin University of China (GSAI) and Ant Group (inclusionAI) · China
Diffusion language models: text is written in blocks and then refined rather than word by word, which speeds up generation. LLaDA2.2 can edit what it has written and targets agents. LLaDA-Image is a separate product.
- Fast generation of code and text
- Agent scenarios with long context
- Research into alternatives to standard LLMs
- Sizes
- 8B – 100B (MoE)
- Hardware
- from: 1 GPU
CodeGGUF2025–2026
JetBrains · Czech Republic
JetBrains models for fast code autocompletion. Mellum2 (12B, 2.5B active) is already a full assistant: it writes and edits code, calls tools and reasons. Version 2.1 spent most of its training on reinforcement learning in real repositories with a shell and tests: 47 on SWE-bench Verified, self-reported by JetBrains.
- Fast code autocompletion on your own server
- A developer assistant that does not send code to the cloud
- Fine-tuning on the company's code
- Sizes
- 4B – 12B-A2.5B
- Hardware
- from: Laptop
TextGGUF2025–2026
Xiaomi · China
Xiaomi models for reasoning and agents: from the compact MiMo-7B to MiMo-V2.6-Pro with 1.02 trillion parameters. The larger versions understand text, images, video and audio, with a 1M token context. Languages: English and Chinese.
- Logic and calculation tasks
- Agents with tools
- Help for developers
- Sizes
- 7B – 1,02T-A42B
- Hardware
- from: Laptop
Text2024–2026
OpenBMB (ModelBest and Tsinghua University) · China
Compact text models that run directly on a device: laptop, phone or mini PC. The 1B and 2B MiniCPM5 models focus on tool calling and long context.
- A local chat assistant without the cloud
- Data extraction and text classification
- Tool calling and simple agents on low-end hardware
- Sizes
- 0.5B – 8B
- Hardware
- from: Laptop
Search and RAGGGUF2026
Perplexity · USA
Embeddings from the Perplexity search service. Some versions take into account the context of the whole document, not just a single fragment.
- Search across large document collections
- RAG that accounts for document context
- Website and catalog search
- Sizes
- 0.6B – 9B
- Hardware
- from: Laptop
Acoustic monitoring2019–2026
Google · USA
A tiny model covering 521 sound events, including alarms, breaking glass and screams: it fits on a microcontroller and runs without a GPU. A microphone can capture people voices, which is personal data, so check the procedure with a lawyer.
- Fast sound labelling right on the device
- Detecting alarm sounds on a site
- Picking interesting fragments out of a continuous recording
- Sizes
- 3.7M
- Hardware
- from: Laptop
TextRU2026
SberDevices (ai-forever) · Russia
A Russian and English research prototype: the model writes text in blocks at once (diffusion) rather than word by word, which speeds up responses. The authors do not recommend it for production systems.
- Experiments with faster generation
- Fine-tuning small models for your own tasks
- Research
- Sizes
- 0.6B – 4B
- Hardware
- from: Laptop
Drawings and CAD2026
FusionBrain Lab, Lomonosov Moscow State University and Innopolis University · Russia
It writes CadQuery code from images and a point cloud, then edits its own output over several rounds, comparing the resulting shape with the original. This is a draft model: an engineer checks and signs off the document.
- Recovering a parametric model from photos and a scan of a part
- Step-by-step refinement of a draft CAD model
- Digitising parts that have no drawings
- Sizes
- 2.2B
- Hardware
- from: Laptop
Search and RAGRU2024–2026
Sber (SberDevices) · Russia
Sber embeddings built for Russian: according to the developers, among the best on Russian-language search benchmarks. FRIDA is compact, Giga-Embeddings is more powerful.
- Search across Russian-language documents
- RAG for chatbots in Russian
- Classifying requests and reviews
- Sizes
- 480M – 10B-A1.8B
- Hardware
- from: Laptop
TextOllama2023–2026
Zhipu AI (Z.ai) · China
One of the oldest Chinese open lines: from ChatGLM-6B to GLM-5.3. Strong at agentic tasks and programming; GLM-5.3-Flash understands images and is released under MIT.
- Corporate chat assistant
- Agents for routine office tasks
- Help for developers
- Sizes
- 1.5B – 744B-A40B
- Hardware
- from: Laptop
Text2024–2026
Tencent · China
Tencent language models: from small 0.5B–7B to Hy4-preview with 770 billion parameters. Since 2026 the line has been renamed Hy, and new versions are released under Apache 2.0.
- Corporate assistant
- Translation and multilingual texts
- Agents with tools
- Sizes
- 0.5B – 770B-A49B
- Hardware
- from: Laptop
TextOllama2024–2026
IBM · USA
IBM enterprise models with transparent training data and ISO 42001 certification. Granite 4 is a memory-efficient Mamba and Transformer hybrid.
- Answers based on internal documents (RAG)
- Tool calling and agent work
- Data extraction and classification
- Sizes
- 350M – 34B
- Hardware
- from: Laptop
Acoustic monitoring2021–2026
MLCommons · USA
A reference autoencoder for finding anomalies in machine sound: it learns only from recordings of a healthy unit and flags deviations. The system does not make a diagnosis, it gives you a reason to check the unit before it fails.
- Listening to a machine tool, pump or conveyor
- Flagging deviations from a unit usual noise
- Training on your own recordings of healthy equipment
- Sizes
- a tiny fully connected autoencoder sized for microcontrollers
- Hardware
- from: Laptop
Speech to textGGUF2024–2026
Moonshine AI (Useful Sensors) · USA
Very small and fast speech recognition models for phones, tablets and embedded devices. Version 2 streams, producing text while the person is still speaking.
- Voice control of devices
- Offline recognition on a phone
- Live subtitles
- Sizes
- 27M – 245M
- Hardware
- from: Laptop
TextOllama2026
Meta Superintelligence Labs · USA
An open Meta model for agents on affordable hardware: distilled from the closed Muse Spark, understands text and images, trained on 100+ languages.
- Agents with tool calling
- Analysis of screenshots, charts and documents
- Multilingual assistant
- Sizes
- 30B
- Hardware
- from: 1 GPU
Computer-use agents2025–2026
XLANG Lab (University of Hong Kong) · China
Fully open desktop agents: weights, data and training code. They work on Windows, macOS and Linux; the latest Qwen-CUA controls a computer with ordinary clicks and keystrokes.
- Working in desktop software without an API
- Moving data between systems
- Running user scenarios for tests
- Sizes
- 7B – about 400B (MoE)
- Hardware
- from: 1 GPU
TextRUOllama2023–2026
Alibaba · China
A family of language models with strong Russian language support, from small versions for a laptop to a flagship on par with commercial APIs.
- Chatbot and knowledge-base assistant
- Replies to emails and customer requests
- Document parsing and classification
- Sizes
- 0,6B – 2,4T-A95B
- Hardware
- from: Laptop
TextOllama2024–2026
LG AI Research · South Korea
Korean-English models from LG. Most of the line is non-commercial, but the flagship K-EXAONE 2.0 with 750 billion parameters is released under Apache 2.0.
- Corporate assistant
- Working with Korean and English texts
- Analysis of documents and images (4.5)
- Sizes
- 1.2B – 750B-A37B
- Hardware
- from: Laptop
Computer-use agentsGGUF2025–2026
Microsoft · USA
Small Microsoft models for working in the browser: they look at the page and click, type and scroll. Designed to run directly on a work computer without the cloud.
- Filling in web forms and applications
- Collecting data from web portals without an API
- Checking websites against scenarios
- Sizes
- 4B – 27B
- Hardware
- from: Laptop
TextGGUF2026
Thinking Machines Lab · USA
Flagship open models from Mira Murati's lab: they take text, images and audio. Large MoE models that need several GPUs.
- Flagship-level corporate assistant
- Analysis of documents, images and audio
- Programming help
- Sizes
- 276B-A12B, 975B-A41B
- Hardware
- from: Cluster
RerankersGGUF2024–2026
Jina AI · Germany
Strong multilingual rerankers; m0 also ranks pages as images (scans, slides). The latest versions are open for non-commercial use only.
- Refining search results before a chatbot answers
- Sorting retrieved PDF pages and slides
- Catalog and knowledge base search
- Sizes
- 33M – 2.4B
- Hardware
- from: Laptop
Code2025–2026
Kwaipilot (Kuaishou) · China
Kuaishou models for agentic development, trained to solve real tasks in repositories. KAT-Coder-V2.5-Dev (35B, 3B active) is the open version of their closed flagship.
- An agent that fixes tasks in the repository
- Code generation and refactoring
- Automating routine development tasks
- Sizes
- 32B – 72B, 35B-A3B
- Hardware
- from: 1 GPU
CodeGGUF2026
Poolside · USA
Models for agentic programming: they edit code in a repository on their own. The small XS runs on a Mac with 36 GB of memory; S 2.1 has a 1M-token context.
- Coding agent for in-house development
- Bug fixing and code improvements
- Working with large codebases
- Sizes
- 33B-A3B – 225B-A23B
- Hardware
- from: 1 GPU
TextRUOllama2023–2026
Mistral AI · France
European models focused on speed. Mixtral was one of the first open mixture-of-experts models; there are versions for images (Pixtral, Medium 3.5), Lean proofs and moderation (Shieldstral).
- Fast chat responses
- Data extraction from text
- Translation and multilingual work
- Sizes
- 3B – 675B
- Hardware
- from: Laptop
Rerankers2026
Tencent · China
A pair of small Tencent models based on Qwen3 that pick the right skill for an AI agent for a given request: the embedding model finds candidates, the reranker chooses the best one.
- Choosing a tool or skill for an AI agent
- Routing requests between bot scenarios
- Search across a catalog of internal tools
- Sizes
- 0.6B
- Hardware
- from: Laptop
TextOllama2023–2026
Upstage · South Korea
Models from Korea's Upstage. Solar Open 2 is built for office document work: 250 billion parameters, 15 billion active; languages are English, Korean and Japanese.
- Working with office documents
- Agents for routine tasks
- Help for developers
- Sizes
- 10.7B – 250B-A15B
- Hardware
- from: 1 GPU
Math and reasoningOllama2025–2026
Open Thoughts (Stanford, Berkeley and other universities) · USA
Fully open reasoning models: both weights and training data are published. Newer OpenThinkerAgent versions can carry out multi-step tasks.
- Calculations and formula checks
- Complex analytics with step-by-step breakdowns
- Checking the logic of internal policies
- Sizes
- 1.5B – 32B
- Hardware
- from: Laptop
TextRUOllama2023–2026
Cognitive Computations (Eric Hartford) · USA
Uncensored fine-tunes of Llama, Mistral, Qwen and others that fulfill almost any request. Filtering and moderation are fully on the deployer; do not show it to customers without your own filter.
- Assistant that does not refuse legal but sensitive topics
- Internal tools under a strict system prompt
- Role-play and creative scenarios
- Sizes
- 0.5B – 405B
- Hardware
- from: Laptop
Computer-use agentsGGUF2025–2026
H Company · France
A French model family for controlling a browser and computer: precisely finds the right element on screen and handles multi-step tasks. The latest Holo3 and 3.1 are open under Apache 2.0.
- Working in web portals and legacy software without an API
- Filling in forms and applications
- Testing interfaces against scenarios
- Sizes
- 0.8B – 235B-A22B
- Hardware
- from: Laptop
Search and RAGRUGGUF2023–2026
Jina AI · Germany
Strong multilingual embeddings with long context; v5-omni understands text, images and audio. Recent versions are open for non-commercial use only.
- Search across documents in many languages
- Search across images and scans
- Classification and clustering
- Sizes
- 33M – 3.8B
- Hardware
- from: Laptop
Forecasting2024–2026
THUML, Tsinghua University · China
A compact forecasting foundation model from the Tsinghua lab: trained on a large set of diverse series and fine-tunable on your own data.
- Forecasting demand and load
- Forecasting sensor readings on the shop floor
- Fine-tuning forecasts on your own history
- Sizes
- 84M (timer-base)
- Hardware
- from: Laptop
Drawings and CADGGUF2026
Autodesk Research · USA
A model from Autodesk Research: from eight rendered views of a part it writes an executable CadQuery program. It was trained on synthetic data only. It produces a draft; an engineer checks and signs off the document.
- Getting a parametric model from a set of views of a part
- Converting finished geometry into editable code
- A draft model for a design office
- Sizes
- 2B (based on Qwen3-VL-2B)
- Hardware
- from: Laptop
TextGGUF2025–2026
ServiceNow · USA
ServiceNow 15B models with step-by-step reasoning that fit on a single GPU. From version 1.5 they also understand images and are good at calling tools.
- A reasoning assistant for internal services
- Tool calling and enterprise agents
- Analysing screenshots and documents with images
- Sizes
- 5B – 15B
- Hardware
- from: Laptop
TextRUGGUF2025–2026
Arcee AI · USA
An American family of MoE models trained from scratch: Nano, Mini and Large. Trinity-Large-Thinking (398B) reasons before answering.
- Agents with tool calling
- Reasoning tasks
- Corporate assistant on your own servers
- Sizes
- 6B – 398B-A13B
- Hardware
- from: Laptop
Search and RAGRUOllama2024–2026
IBM · USA
Lightweight IBM embeddings for enterprise search, trained on data with clear rights. R2, released in 2026, became multilingual.
- Search across corporate documents
- RAG on a regular server without a GPU
- Reranking results
- Sizes
- 30M – 311M
- Hardware
- from: Laptop
Moderation and safetyOllama2024–2026
IBM · USA
IBM judge models: they catch harm, profanity and jailbreak attempts, and in RAG and agents check whether an answer is grounded in the documents. You can state your own rule in words.
- Checking bot requests and replies
- Finding made-up facts in knowledge-base answers
- Checking your own rules written as text
- Sizes
- 38M – 8B
- Hardware
- from: Laptop
Text to speech2025–2026
Kyutai · France
Streaming speech recognition and synthesis models from the makers of Moshi: they start speaking and transcribing without waiting for the end of a phrase. Pocket TTS (100M) runs on a CPU. English, French and a few other European languages, no Russian.
- Streaming speech transcription for voice bots
- Voicing replies with minimal delay
- Speech synthesis on a server without a GPU (Pocket TTS)
- Sizes
- 100M (Pocket TTS) – 2.6B
- Hardware
- from: Laptop
Search and RAGGGUF2025–2026
Octen · USA / Singapore
Qwen3-Embedding models fine-tuned by the startup Octen for search in legal, financial and medical texts. As of January 2026 the 8B version topped the RTEB leaderboard.
- Search across contracts and case law
- Search across financial reports
- Search across long documents up to 32K tokens
- Sizes
- 0.6B – 8B
- Hardware
- from: Laptop
Fact-checking and judgesRU2025–2026
SberDevices (ai-forever) · Russia
Judge models that evaluate other AI models' answers in Russian: they score against a given criterion and explain the score in text.
- Automatic quality checks of Russian chatbot answers
- Comparing several models before choosing one
- Checking answers after fine-tuning
- Sizes
- 4B – 32B
- Hardware
- from: Laptop
CodeOllama2025–2026
Essential AI · USA
An 8B model trained from scratch by the company of one of the authors of the transformer architecture. Strong at code and technical tasks; version 1.5 handles context up to 160K tokens.
- Writing and fixing code
- A developer agent on a single GPU
- Solving technical and scientific problems
- Sizes
- 8B
- Hardware
- from: Laptop
Computer-use agents2024–2026
Show Lab (National University of Singapore) · Singapore
A lightweight model for working with interfaces: finds buttons and fields by description and performs actions on the web and on a phone. ShowUI-π can drag with the mouse.
- Clicking and filling in forms from a task description
- Web UI autotests
- An assistant on a low-end computer without the cloud
- Sizes
- 2B (ShowUI), about 500M (ShowUI-π)
- Hardware
- from: Laptop
ForecastingGGUF2025–2026
Datadog · USA
A Datadog forecasting model trained on server and application metrics. Especially strong for IT monitoring: load, latency, errors.
- Server load forecasting
- Anomaly detection in metrics
- Capacity planning
- Sizes
- 4M – 2.5B
- Hardware
- from: Laptop
Acoustic monitoring2023–2026
Xiaomi · China
Compact sound-labelling models from 5.5M to 86M, with ONNX and INT8 builds that run on an ordinary CPU and on a board next to the equipment. The system does not make a diagnosis, it gives you a reason to check the unit before it fails.
- Labelling sound on modest hardware and on site
- Detecting alarm sounds and abnormal noise
- Fine-tuning for your own set of equipment sounds
- Sizes
- 5.5M - 86M
- Hardware
- from: Laptop
Voice: speakers and soundRU2026
FireRedTeam (Xiaohongshu) · China
A speech and sound event detector: tells apart speech, singing and music. In a 102-language test (the FLEURS set, which includes Russian) it beat Silero VAD and TEN VAD. Has a streaming mode.
- Cutting recordings before speech recognition
- Separating speech from music and singing in broadcasts and videos
- Speech detection in voice bots
- Sizes
- compact, exact size not stated
- Hardware
- from: Laptop
Math and reasoningGGUF2025–2026
Princeton University · USA
Open models for formal proofs in Lean 4 from Princeton. The new Goedel-Code-Prover proves program correctness.
- Formal verification of mathematical workings
- Verifying code correctness
- Training
- Sizes
- 7B – 32B
- Hardware
- from: Laptop
Search and RAGRU2026
Microsoft · USA
Microsoft's 2026 multilingual embeddings with context up to 32K tokens; Russian is on the language list. The 270M and 0.6B versions run on a regular server, 27B is the most accurate.
- Multilingual knowledge base search
- Picking passages for RAG
- Search across long documents
- Sizes
- 270M – 27B
- Hardware
- from: Laptop
Drawings and CADGGUF2025–2026
Fudan University and StepFun · China
A family of models that produce SVG both from an image and from a text description, from a simple icon to a detailed illustration. For precise drawings the result has to be checked against the original.
- Vectorising images and symbols into SVG
- Creating icons and pictograms from a text description
- Preparing vector graphics for web and print
- Sizes
- 3B – 8B
- Hardware
- from: 1 GPU
Computer-use agents2025–2026
Alibaba (Tongyi Lab, X-PLUG) · China
Models for controlling phones and computers from the Mobile-Agent project: they work with Android, Windows, macOS and the browser; version 1.5 has a reasoning mode.
- Automating actions in mobile apps
- Working in desktop software without an API
- Testing apps against scenarios
- Sizes
- 2B – 32B
- Hardware
- from: Laptop
CodeGGUF2026
IQuest Research · China
A family of coding models with standard and reasoning versions, including a Loop variant that runs through its layers a second time. Sizes from 7B to 40B.
- Writing and refining code
- Solving tasks with step-by-step reasoning
- Agentic work with a repository
- Sizes
- 7B – 40B
- Hardware
- from: Laptop
Voice assistantsGGUF2025–2026
OpenBMB (ModelBest, Tsinghua University) · China
A small model that sees, hears and replies by voice in real time, and can clone a voice. Voice dialogue in English and Chinese, text in 30+ languages.
- Voice assistant on your own server
- Analyzing videos and documents
- Voice answers about a camera image
- Sizes
- 8B – 9B
- Hardware
- from: Laptop
Voice assistants2024–2026
Kyutai · France
A voice assistant that listens and speaks at the same time, with no delay for recognition and synthesis. Hibiki does simultaneous speech-to-speech translation between several European languages.
- Real-time voice conversation partner
- Simultaneous speech translation
- Zero-latency voice interfaces
- Sizes
- 2B – 7B
- Hardware
- from: Laptop
CodeOllama2024–2026
Alibaba (Qwen team) · China
The broadest open coding family: from 0.5B for autocompletion to 480B for agents. Qwen3-Coder-Next (80B, 3B active) works as a developer agent on a single GPU.
- Code autocompletion in the editor
- An agent that edits code in the repository on its own
- Writing and refining scripts, SQL and integrations
- Sizes
- 0.5B – 480B-A35B
- Hardware
- from: Laptop
Code2026
Ai2 (Allen Institute for AI) · USA
Fully open developer agents from Ai2: weights, data and training recipe are all public. Designed so a company can cheaply fine-tune the agent on its own repository.
- An agent for fixing issues in code
- Fine-tuning the agent on an internal repository
- Automating small edits and tests
- Sizes
- 8B – 32B
- Hardware
- from: Laptop
Computer-use agentsGGUF2026
Meituan · China
Meituan's computer-control agent, trained on a large number of simulated tasks in desktop software. It outputs clicks and keyboard input.
- Working in office and legacy software without an API
- Moving data between systems
- Running test scenarios
- Sizes
- 8B – 32B
- Hardware
- from: 1 GPU
Cybersecurity2025–2026
Cisco (Foundation AI) · USA
Cisco models for information security based on Llama 3.1 8B: analysis of vulnerabilities, threats and incidents. Can be deployed inside your own perimeter.
- Analyzing vulnerability and threat reports
- Helping SOC analysts during incidents
- Mapping threats to MITRE ATT&CK
- Sizes
- 8B
- Hardware
- from: Laptop
TextGGUF2024–2026
Prime Intellect · USA
Models trained in a distributed way on GPUs from around the world. INTELLECT-3 (106B) is further trained with reinforcement learning for math, code and agents.
- Reasoning and math tasks
- Programming help
- Agents with tool calling
- Sizes
- 10B – 106B-A12B
- Hardware
- from: 1 GPU
TextRUOllama2023–2026
Microsoft · USA
Small Microsoft models trained on carefully selected data: strong at logic and math for their modest size. Versions with images and speech are available.
- Assistant on a laptop or your own server
- Reasoning and calculation tasks
- Analysis of images and diagrams (vision versions)
- Sizes
- 1.3B – 42B-A6.6B
- Hardware
- from: Laptop
Search and RAGRUOllama2025–2026
Alibaba (Qwen) · China
Embeddings and rerankers based on Qwen3, among the best open ones for multilingual search, including Russian. VL versions search images, screenshots and video.
- Knowledge base search for RAG
- Reranking results before answering
- Search across scans, slides and screenshots
- Sizes
- 0.6B – 8B
- Hardware
- from: Laptop
Voice: speakers and soundRU2025–2026
Daily (Pipecat) · USA
Uses intonation to tell whether a person has finished a thought or just paused, so a voice bot does not interrupt. Version 3 is 8 MB, runs on a CPU and understands 23 languages, including Russian.
- Voice bot does not interrupt the customer during pauses
- Fast reply when the customer has really finished
- An add-on to a standard speech detector in voice assistants
- Sizes
- 8M (v3) – 580M (v1)
- Hardware
- from: Laptop
Search and RAG2026
Voyage AI (MongoDB) · USA
The only open model in the Voyage 4 line: its vectors are compatible with the paid larger versions, so you can start locally and move to the API later.
- Document search on your own server
- RAG for small knowledge bases
- Finding similar texts
- Sizes
- about 340M
- Hardware
- from: Laptop
Moderation and safety2025
ServiceNow · USA
A guard model that catches both harmful content and attacks on AI (prompt injection, jailbreaks), including when agents use tools.
- Screening chatbot requests for attacks and jailbreaks
- Filtering harmful model answers
- Monitoring the actions of AI agents that use tools
- Sizes
- 8B
- Hardware
- from: Laptop
Computer-use agents2023–2025
Zhipu AI (Z.ai) and Tsinghua University · China
One of the first open models for controlling an interface from a screenshot; its successor, AutoGLM-Phone, works in Android smartphone apps.
- Automating actions in mobile apps
- Working in web interfaces without an API
- Testing apps against scenarios
- Sizes
- 9B – 18B
- Hardware
- from: 1 GPU
CodeRUOllama2025
Mistral AI (with All Hands AI) · France
Mistral models for agentic development: they read the repository, edit files and run commands on their own. The 24B version fits on a single GPU.
- A developer agent that fixes tickets from the tracker
- Extending internal systems from a description
- Automating routine code edits
- Sizes
- 24B – 123B
- Hardware
- from: 1 GPU
Computer-use agentsGGUF2025
Alibaba (Tongyi-MAI) · China
Compact Alibaba models for working in smartphone and computer interfaces: they find elements and complete multi-step tasks. The agent acts on the user's behalf inside their apps, so access rights and limits are set before launch. The follow-up line, Qwen-UI-Agent, was announced in July 2026, but its weights have not been released.
- Automating actions in mobile apps
- Working in software without an API
- UI autotests
- Sizes
- 2B – 8B
- Hardware
- from: Laptop
Tabular dataGGUF2024–2025
Zhejiang University · China
A family for working with tables and databases: it understands data structure, writes parsing code and answers questions about exports.
- Answering questions about tables and data exports
- Automated data analysis with generated code
- A helper for BI and internal reporting
- Sizes
- 7B – 72B
- Hardware
- from: 1 GPU
TextOllama2025
Deep Cogito · USA
Fine-tuned Llama, Qwen and DeepSeek models with a hybrid mode: answer immediately or reason first. The 671B v2.1 flagship spends noticeably fewer tokens on reasoning than DeepSeek R1.
- A chat assistant with a reasoning mode
- Writing code and calling tools
- Answering complex questions about documents
- Sizes
- 3B – 671B
- Hardware
- from: Laptop
TextOllama2023–2025
Nous Research · USA
Nous Research fine-tunes on top of Llama, Mistral, Qwen and Seed-OSS. Valued for precise instruction following, function calling and strict JSON output; they refuse less often than the originals; Hermes 4 has a reasoning mode.
- Agents that call functions and APIs
- Data extraction in strict JSON format
- Assistant with flexible role and tone settings
- Sizes
- 3B – 405B
- Hardware
- from: Laptop
Acoustic monitoring2025
Carnegie Mellon University and co-authors (the ESPnet project) · USA
A fully open reproduction of BEATs: code, training recipes and weights are all published. Teams pick it when they need a transparent base for their own sound model. The system does not make a diagnosis, it gives you a reason to check the unit before it fails.
- A transparent base for your own audio model
- Fine-tuning on your own equipment recordings
- Labelling sound events
- Sizes
- base and large versions, exact sizes are listed on the model cards
- Hardware
- from: Laptop
Drawings and CAD2025
Kingsoft · China
It reads a flowchart from an image and first writes out its structure as JSON - blocks, links and branching conditions - and only then answers a question about the chart. The parse still has to be checked against the original.
- Converting a flowchart image into a JSON structure
- Answering questions about a process diagram
- Digitising procedures and diagrams from presentations
- Sizes
- 32B
- Hardware
- from: 1 GPU
Voice: speakers and soundRU2020–2025
Silero · Russia
The most popular open speech detector: tells voice apart from silence and noise. Processes an audio chunk in under a millisecond on a single CPU core; trained on recordings in more than 6,000 languages.
- Cutting calls and recordings before speech recognition
- Detecting when the customer is speaking in a voice bot
- Filtering out silence and noise to save on transcription
- Sizes
- about 2 MB
- Hardware
- from: Laptop
Rerankers2025
ZeroEntropy · USA
Rerankers built on Qwen3. The model card lists the target domains — finance, law, code, medicine, science; the stated language is English.
- Refining results before an AI assistant answers
- Sorting search results across contracts and reports
- Search across technical and scientific documentation
- Sizes
- zerank-2 — 4B (based on Qwen3-4B), plus a smaller "small" version
- Hardware
- from: Laptop
Recommenders2025
MiniOneRec open project · China
A small open generative recommender at 1.5B with published checkpoints and a full training pipeline. Handy as a learning and starter option; a real store needs fine-tuning on its own orders.
- A quick recommendation prototype on your own data
- Testing ideas before running a large model
- Training a team to work with generative recommendations
- Sizes
- 1.5B
- Hardware
- from: Laptop
Visual document search2025
Illuin Technology, EPFL, CentraleSupélec · France
A compact (250M) model for searching document pages as images. According to the authors, it matches models 10 times larger and runs without a GPU.
- Search across scans and PDFs on a modest server
- Indexing document archives
- Search across slides and manuals
- Sizes
- 250M
- Hardware
- from: Laptop
Search and RAGRUOllama2024–2025
Mixedbread · Germany
Embeddings and rerankers from Germany's Mixedbread. mxbai-embed-large is one of the most downloaded English search models; the v2 rerankers cover 100+ languages, including Russian.
- Search across a knowledge base
- Reranking results before a bot answers
- Product catalog search
- Sizes
- 17M – 1.5B
- Hardware
- from: Laptop
Search and RAGOllama2023–2025
BAAI (Beijing Academy of Artificial Intelligence) · China
Some of the most popular embeddings for search and RAG. The main v1.5 versions target English and Chinese; for Russian, BAAI has a separate model, bge-m3.
- Search across English-language documents
- Picking passages for chatbot answers (RAG)
- Code search (bge-code)
- Sizes
- 24M – 9B
- Hardware
- from: Laptop
Text to SQLGGUF2024–2025
Prem AI · UK
A text-to-SQL model of just 1B parameters, designed to run locally so the database never leaves for external services.
- Local translation of questions into SQL with no internet access
- Query hints on modest hardware
- Embedding into internal analytics tools
- Sizes
- 1B
- Hardware
- from: Laptop
Drawings and CAD2025
Manycore Research · China
From a point cloud of a room or a walkthrough video it builds a structured plan: walls, doors, windows and objects with coordinates and sizes. It is a draft survey; final areas are confirmed by a specialist.
- Building a floor plan from a 3D scan of a space
- Extracting walls, doors and windows with dimensions
- Draft measurements before a renovation or a move
- Sizes
- 0.5B - 1B
- Hardware
- from: Laptop
TextOllama2025
OpenAI · USA
OpenAI's first open models since GPT-2. Reasoning and tool calling; the smaller version fits on a single GPU.
- AI agent that calls internal systems
- Answers based on internal policies
- Drafts of emails and reports
- Sizes
- 20B, 120B
- Hardware
- from: 1 GPU
Text2024–2025
xAI · USA
xAI publishes the weights of previous Grok generations. The models are very large and need a GPU cluster, so in practice they are rarely run.
- Research on large models
- Assistant on your own infrastructure
- Text generation and analysis
- Sizes
- 314B (Grok-1), Grok-2 is larger
- Hardware
- from: Cluster
Math and reasoningGGUF2025
Moonshot AI and Project Numina · China, France
Models for formal proofs in Lean 4 from Moonshot AI (Kimi) and Numina. Small versions from 0.6B run on a laptop.
- Formal verification of mathematical workings
- Translating a problem from plain language into Lean
- Training and olympiad preparation
- Sizes
- 0.6B – 72B
- Hardware
- from: Laptop
TextGGUF2025
ByteDance · China
An open ByteDance 36B model with up to 512K tokens of context and an adjustable thinking budget. Fits on a single powerful GPU.
- Analysis of long documents
- Agents with tools
- Corporate assistant
- Sizes
- 36B
- Hardware
- from: 1 GPU
Visual document search2024–2025
Illuin Technology (ViDoRe team) · France
Searches PDFs and scans as images: pages do not need to be OCR'd first, the model finds the right one for a question directly, including tables and charts. Trained on English.
- Search across scans, presentations and PDFs
- RAG over documents with tables and charts
- Search across technical documentation
- Sizes
- 256M – 3B
- Hardware
- from: Laptop
Fact-checking and judges2024–2025
OpenCompass (Shanghai AI Laboratory) · China
A line of judges from the team behind open model benchmarks: they score answers and check them against a reference. The judge itself makes mistakes and does not replace manual review on important tasks.
- Scoring model answers against set criteria
- Checking an answer against a reference solution
- Comparing several models on your own data
- Sizes
- 1.5B – 32B
- Hardware
- from: Laptop
Math and reasoningOllama2025
Agentica (Berkeley, Sky Computing Lab) and Together AI · USA
Small models fine-tuned with reinforcement learning: DeepScaleR (1.5B) solves olympiad maths, DeepCoder writes code, DeepSWE works as a developer agent. Recipes and data are open.
- Solving maths problems with step-by-step working
- Generating and checking code
- An agent for fixing bugs in a repository
- Sizes
- 1.5B – 32B
- Hardware
- from: Laptop
Voice: speakers and sound2025
Agora (TEN project) · USA / China
A lightweight speech detector for real-time voice assistants: it notices the start and end of a phrase faster than Silero VAD. Runs on servers, phones and in the browser.
- Zero-lag speech detection in a voice bot
- Fast assistant response at the end of a phrase
- Use in mobile apps and the browser
- Sizes
- very small, the library is smaller than Silero VAD
- Hardware
- from: Laptop
Math and reasoning2025
NVIDIA · USA
NVIDIA models for maths and reasoning based on Qwen. AceReason was fine-tuned with reinforcement learning first on maths, then on code.
- Calculations and formula checks
- Complex analytics with step-by-step breakdowns
- Working through programming problems
- Sizes
- 1.5B – 72B
- Hardware
- from: Laptop
Drawings and CAD2025
AIRI, Lomonosov Moscow State University and Innopolis University · Russia
A successor to CAD-Recode: it builds a CAD model as CadQuery code from three kinds of input at once - a point cloud, an image and a text description. It produces a draft; an engineer checks and signs off the document.
- Building a CAD model from a photo of a part
- Recovering a model from a 3D scan
- A draft part from a short text description
- Sizes
- 2.2B (based on Qwen2-VL)
- Hardware
- from: Laptop
Cybersecurity2022–2025
NVIDIA · USA
An open NVIDIA bundle: a card-transaction graph plus a classifier, a profile of normal user behaviour, log parsing and root-cause search. A human makes the decision on a disputed operation; automatic blocking without review is not acceptable.
- Highlighting suspicious card operations taking participant links into account
- Tracking deviations of an account from its usual behaviour
- Parsing logs and finding the root cause of an incident
- Sizes
- from a few megabytes (tree-based models) up to BERT-based classifiers
- Hardware
- from: 1 GPU
CybersecurityGGUF2023–2025
Clouditera · China
A Chinese open family for cybersecurity: reviewing vulnerabilities, analysing logs and traffic, explaining commands and scripts.
- Reviewing vulnerabilities and drafting fix recommendations
- Analysing logs and reconstructing an attack chain
- Explaining suspicious commands and scripts
- Sizes
- 1.5B – 14B
- Hardware
- from: Laptop
Fact-checking and judges2024–2025
Skywork (Kunlun Tech) · China
Reward models: they score how good a language model's answer is for the user. Used for fine-tuning your own models and picking the best of several answers.
- Choosing the best of several bot answers
- Scoring answer quality during model fine-tuning
- Comparing models before rollout
- Sizes
- 0.6B – 27B
- Hardware
- from: Laptop
Text to SQLGGUF2025
IDEA Research · China
A text-to-SQL model trained with reinforcement learning: it works through the schema and the conditions step by step before producing a query.
- Database queries for questions with several conditions
- Reviewing and fixing other people SQL queries
- An analyst helper inside a BI system
- Sizes
- 3B – 14B
- Hardware
- from: Laptop
Drawings and CAD2024–2025
DFKI and RPTU Kaiserslautern · Germany
It builds a sequence of CAD modelling commands straight from a text description of a part, from a beginner phrase to a designer-level wording. The output is a draft; an engineer checks and signs off the document.
- A draft part from a text description
- Quickly trying out shape options at an early stage
- Filling a library of standard parts
- Sizes
- size not stated on the model card
- Hardware
- from: 1 GPU
CybersecurityGGUF2025
Trendyol · Turkey
Security models from a large Turkish marketplace, published in GGUF format: reviewing alerts and incidents, English and Turkish.
- Reviewing alerts and first-pass incident assessment
- Explaining suspicious activity in reports
- Helping the on-duty shift of a monitoring centre
- Sizes
- 32B и 70B
- Hardware
- from: 1 GPU
Text to SQL2025
Snowflake · USA
A Snowflake model for turning questions into SQL, trained with reinforcement learning by checking query results. The open 7B version is based on Qwen2.5-Coder.
- Plain-language questions to a data warehouse
- Generating SQL for reports and dashboards
- Checking and fixing analysts' queries
- Sizes
- 7B
- Hardware
- from: Laptop
Image generationGGUF2025
ByteDance Seed · China
A unified model that understands images, generates them and edits them in a conversation. Similar to how images work in ChatGPT.
- Photo editing in a conversation
- Answering questions about an image
- Image generation with explanations
- Sizes
- 14B-A7B
- Hardware
- from: 1 GPU
Rerankers2023–2025
Stanford NLP, later Answer.AI and LightOn · USA and France
A different search principle: every word of the question is compared with every word of the document, not the two texts as a whole. The index is heavier than with ordinary embeddings. The model cards list English.
- Search across a knowledge base of long documents
- Reordering retrieved passages
- Search across policies and technical documentation
- Sizes
- about 33M – 150M
- Hardware
- from: Laptop
TextOllama2025
DeepSeek · China
A reasoning model that thinks step by step before answering. Strong at calculations, logic and code; compact distilled versions are available.
- Complex calculations and logic checks
- Analysis of contracts and internal policies
- Help for developers
- Sizes
- 1,5B – 671B
- Hardware
- from: Laptop
Acoustic monitoring2024–2025
Shanghai Jiao Tong University and Peng Cheng Laboratory · China
A self-supervised sound understanding model that is markedly cheaper to train on your own data than its predecessors. The system does not make a diagnosis, it gives you a reason to check the unit before it fails.
- Training your own audio model on modest hardware
- Audio features for spotting abnormal operating modes
- Labelling sound events on a site
- Sizes
- 90M (base) and 309M (large)
- Hardware
- from: Laptop
CodeRU2025
MTS AI (MWS AI) · Russia
A small coding assistant from MTS AI that understands requests in Russian. Runs locally, with plugins for VS Code and JetBrains.
- Code suggestions and completion in the editor
- Code explanations in Russian
- Drafts of tests and documentation
- Sizes
- 1.5B
- Hardware
- from: Laptop
Fact-checking and judgesRU2024–2025
NVIDIA · USA
Large NVIDIA scorers for selecting and fine-tuning answers. The multilingual GenRM version lists Russian among its languages. The scorer itself makes mistakes and does not replace manual review on important tasks.
- Choosing the best of several candidate answers
- Preparing data to fine-tune your own model
- Scoring assistant answers in Russian and other languages
- Sizes
- 49B, 70B and 340B
- Hardware
- from: Cluster
Text analysisRU2023–2025
deepvk (VK) · Russia
Russian encoders from the VK team: RuModernBERT reads long texts, USER produces vectors for search, GeRaCl classifies texts by topic without training.
- Classifying requests without labeled data
- Knowledge base search in Russian
- Analyzing long contracts
- Sizes
- 35M – 360M
- Hardware
- from: Laptop
CodeGGUF2025
ByteDance Seed · China
A compact 8B coding model from ByteDance in base, instruct and reasoning versions. Its training data was selected by the model itself, with almost no hand-written rules.
- Code autocompletion and generation
- Solving algorithmic problems
- A base for fine-tuning on your own stack
- Sizes
- 8B
- Hardware
- from: Laptop
Forecasting2025
THUML, Tsinghua University · China
A forecasting model that returns a set of possible scenarios rather than a single line — useful when you need a range for demand or load, not one number.
- Forecasting demand with a range of values
- Planning stock while accounting for spread
- Forecasting load on services and staff
- Sizes
- 128M (sundial-base)
- Hardware
- from: Laptop
Math and reasoningGGUF2024–2025
DeepSeek · China
DeepSeek models for formal proofs in Lean 4: the proof is checked by a program, not a person. A narrow tool for mathematicians and engineers.
- Formal verification of mathematical workings
- Verifying algorithm correctness
- Training and olympiad preparation
- Sizes
- 7B – 671B
- Hardware
- from: Laptop
TextOllama2023–2025
Meta · USA
The models that started mass open source in AI. A huge ecosystem of fine-tuned versions and tools.
- Assistant for employees
- Summaries of meetings and documents
- Base for industry-specific fine-tuning
- Sizes
- 1B – 405B
- Hardware
- from: Laptop
Moderation and safety2024–2025
Meta · USA
Tiny classifiers that catch attempts to hack a bot: prompt injections and rule bypassing. The 86M version is multilingual, 22M is English only.
- Protecting a bot from prompt injections
- Checking emails and documents that reach an AI agent
- Fast filter in front of a large model
- Sizes
- 22M – 86M
- Hardware
- from: Laptop
Computer-use agentsGGUF2025
ByteDance Seed · China
A model that looks at a screenshot and controls the mouse and keyboard itself: clicks, fills in fields, navigates menus. The first generation and 1.5-7B are open; UI-TARS-2 weights were not released.
- Working in legacy software without an API
- Filling in forms and moving data between systems
- UI autotests from plain-language scenarios
- Sizes
- 2B – 72B
- Hardware
- from: Laptop
CybersecurityGGUF2023–2025
Kindo · USA
One of the best-known open families for security and DevSecOps work: reviewing code for weaknesses, test scenarios, explaining attacks.
- Finding weak spots in code and configurations
- Reviewing incidents and explaining attack techniques
- Drafting scripts and procedures for the security team
- Sizes
- 7B – 70B
- Hardware
- from: Laptop
Text to SQL2025
Alibaba · China
Alibaba models for turning questions into SQL, based on Qwen2.5-Coder. They work with different SQL dialects; a small 3B version suits modest hardware.
- Plain-language database questions
- Queries for different databases (PostgreSQL, MySQL, SQLite)
- Automating routine reports
- Sizes
- 3B – 32B
- Hardware
- from: Laptop
Drawings and CAD2024–2025
University of Luxembourg (SnT) and Artec3D · Luxembourg
Turns a point cloud of a part into executable CadQuery Python code, that is, a parametric 3D model you can open and edit in CAD. The output is a draft: an engineer checks and signs off the document.
- Recovering a parametric model from a 3D scan of a part
- Converting a point cloud into editable CadQuery code
- Preparing a draft model for a design engineer to finish
- Sizes
- 1.5B (Qwen2-1.5B plus a single projection layer)
- Hardware
- from: Laptop
Search and RAGRUOllama2024–2025
Nomic AI · USA
Fully open embeddings, with weights, data and training code. v2 is multilingual on MoE; there are versions for code and for searching PDF pages.
- Search across documents and a knowledge base
- Code search
- Search across scans and PDFs without text recognition
- Sizes
- 137M – 7B
- Hardware
- from: Laptop
Text to SQL2025
Renmin University of China (RUC) · China
Models for turning questions into SQL, trained on millions of synthetic query examples across different databases. Three sizes for different hardware.
- Database questions without knowing SQL
- Generating queries for reports
- A base for fine-tuning on your own database schema
- Sizes
- 7B – 32B
- Hardware
- from: Laptop
Text to speechGGUF2025
Canopy Labs · USA
Language-model-based speech synthesis with lively intonation and emotional cues. Responds quickly, suitable for voice assistants. Mainly English.
- Real-time voice for an assistant
- Emotional voiceover
- Voice cloning
- Sizes
- 3B
- Hardware
- from: Laptop
Text to speechRUGGUF2023–2025
Rhasspy / Open Home Foundation · USA
Very fast speech synthesis that runs even on a Raspberry Pi. Ready-made voices in 35+ languages, including several Russian ones.
- Voicing notifications and bot replies
- Voice for offline devices
- Voice menus
- Sizes
- about 5M – 30M
- Hardware
- from: Laptop
Image generationGGUF2024–2025
NVIDIA · USA
NVIDIA's fast image model: 4K images in seconds, runs even on a laptop GPU. The Sprint version generates in 1–2 steps.
- Bulk image generation
- High-resolution visuals
- Real-time generation inside apps
- Sizes
- 0.6B – 4.8B
- Hardware
- from: Laptop
Text to speechGGUF2025
Sesame · USA
A conversational speech model that takes the context of the conversation into account and sounds like a real person. English only.
- Voice for a conversational assistant
- Voicing dialogues
- Voice product prototypes
- Sizes
- 1B
- Hardware
- from: Laptop
Drawings and CADGGUF2025
ServiceNow Research, Mila and ETS Montreal · Canada
It turns a raster image into SVG code: instead of tracing pixels the model writes vector commands, so lines and symbols stay editable. It simplifies complex drawings, so the result needs a visual check.
- Vectorising diagrams, symbols and logos from raster images
- Turning scans of simple drawings into editable vectors
- Recovering icons and pictograms from images
- Sizes
- 1B and 8B
- Hardware
- from: Laptop
Computer-use agents2024–2025
Salesforce · USA
Salesforce models for function calling and agents: they pick the right tool and fill in its parameters. Strong on benchmarks, but the license is non-commercial.
- Calling APIs and internal services on user request
- Multi-step agents with several tools
- Comparing approaches before choosing a commercial model
- Sizes
- 1B – 8x22B
- Hardware
- from: Laptop
Text to speechGGUF2024–2025
hexgrad (independent developer) · not disclosed
A tiny speech synthesis model (82M) that sounds on par with large ones. Runs on a regular CPU; English and a few other languages, no Russian.
- Voicing articles and notifications
- Voice for apps without a GPU
- Bulk text voiceover
- Sizes
- 82M
- Hardware
- from: Laptop
Computer-use agentsGGUF2025
Microsoft Research · USA
An agent model that plans actions both in an interface (buttons on screen) and for a robot (arm movements). For now more of a research base than a finished product.
- Pilots in interface control
- Research projects spanning screens and robotics
- Analyzing screenshots with an action plan
- Sizes
- 8B
- Hardware
- from: 1 GPU
Computer-use agents2024–2025
Microsoft · USA
Breaks a screenshot down into buttons, fields and icons with labels so a regular language model can understand and control the screen. It does not click itself; it serves as the agent's eyes.
- Mapping legacy software screens for automation
- Preparing an agent to work in an interface
- Checking that the required elements are on screen
- Sizes
- under 1B (detector + captioning)
- Hardware
- from: Laptop
Computer vision2023–2025
Google · USA
Models that map images and text into a shared space: you can search photos by words and classify images without training. OpenAI's CLIP (2021) is the predecessor.
- Image search by text query
- Automatic catalog labeling and tagging
- Filtering prohibited content
- Sizes
- about 0.2B to 2B
- Hardware
- from: Laptop
Fact-checking and judges2025
Atla · UK
An 8B judge model: it scores another model answer against your criteria and writes a rationale. The judge itself makes mistakes and does not replace manual review on important tasks.
- Scoring chatbot answers against your own criteria
- Comparing two versions of a prompt or model
- Filtering out weak answers before they reach a person
- Sizes
- 8B
- Hardware
- from: 1 GPU
Search and RAGRU2023–2025
Alibaba · China
Alibaba embeddings and rerankers for search: from tiny to 7B based on Qwen2. There is a multilingual mGTE version with long context.
- Semantic search across documents
- Reranking search results
- Clustering and classifying texts
- Sizes
- 33M – 7B
- Hardware
- from: Laptop
Text analysisGGUF2024–2025
Answer.AI and LightOn · USA / France
A modern replacement for classic BERT: faster, reads up to 8 thousand tokens at once. A base for your own classifiers. Trained on English and code; for Russian there is RuModernBERT.
- Classifying requests and documents
- Finding relevant passages in long texts
- Base for your own classifier after fine-tuning
- Sizes
- 150M – 395M
- Hardware
- from: Laptop
Text analysisRUOllama2024–2025
Jina AI · Germany
Small models that turn raw web page HTML into clean Markdown or JSON. Handy for preparing websites for a knowledge base. Non-commercial license only.
- Cleaning website pages for a knowledge base
- Extracting data from pages into JSON
- Preparing texts for RAG
- Sizes
- 0.5B – 1.5B
- Hardware
- from: Laptop
Search and RAGRUOllama2019–2025
UKP Lab (TU Darmstadt), later Hugging Face · Germany
The classic for meaning-based search: small, fast models that run even on a modest server without a GPU. The multilingual versions understand Russian.
- Search across a knowledge base and FAQ
- Finding similar tickets and duplicates
- Grouping reviews and requests by topic
- Sizes
- about 20M – 470M
- Hardware
- from: Laptop
Fact-checking and judges2024
Patronus AI · USA
A small judge: it scores against your criteria and highlights which part of the answer led to that score. The license is non-commercial. The judge itself makes mistakes and does not replace manual review.
- Scoring answers against your criteria with an explanation
- Understanding why a score was lowered
- Bulk review of assistant conversations
- Sizes
- 3.8B (based on Phi-3.5-mini)
- Hardware
- from: Laptop
Visual document search2024
Alibaba (Tongyi Lab) · China
One vector for text, for an image and for a text-image pair: a single model can find a product by photo, a document page by question and an image by description. The card lists English and Chinese.
- Finding a product by photo
- Search across a catalogue of images and cards
- Search across document pages as images
- Sizes
- 2B and 7B
- Hardware
- from: 1 GPU
Search and RAGRUOllama2024
Snowflake · USA
Snowflake embeddings built specifically for search. Version 2.0 is multilingual (Russian is on the language list), handles long texts up to 8K tokens and can compress vectors.
- Search across documents and knowledge bases
- Picking passages for RAG
- Search across reports and internal data
- Sizes
- 22M – 568M
- Hardware
- from: Laptop
TextOllama2024
Nexusflow · USA
Fine-tuned Llama 3 and Qwen 2.5 models from Nexusflow. Athene-V2-Agent is specially trained for function calling and agent scenarios. Commercial use is prohibited.
- Research on agents and function calling
- Comparison with commercial models
- Experiments with a chat assistant
- Sizes
- 70B – 72B
- Hardware
- from: 1 GPU
CodeOllama2024
INF Technology · China
Fully reproducible coding models: along with the weights, the data, its cleaning pipeline and the training recipe are open. Understand English and Chinese.
- Code generation and completion
- Training your own coding model from an open recipe
- A programming assistant on low-end hardware
- Sizes
- 1.5B – 8B
- Hardware
- from: Laptop
Forecasting2024
IBM Research · USA
Tiny forecasting models from IBM: they run on an ordinary CPU and sit next to the business system without a separate GPU server.
- Forecasting sales and warehouse stock
- Forecasting energy use and equipment load
- Fast forecasts right on the company server
- Sizes
- very small: TinyTimeMixers have about 1M parameters
- Hardware
- from: Laptop
Forecasting2024
Auton Lab, Carnegie Mellon University · USA
A foundation model for numeric series: one engine is used for forecasting, anomaly detection, filling gaps and classification.
- Forecasting demand and load
- Detecting anomalies in sensor readings and metrics
- Filling gaps in historical data
- Sizes
- about 40M – 385M
- Hardware
- from: Laptop
Visual document search2024
LightOn · France
A reranker for document pages as images: after a visual search it reorders the found pages by how well they answer the question. The card does not state the languages.
- Refining search results over scans and PDFs
- Selecting pages before an AI assistant answers
- Sorting retrieved slides and reports
- Sizes
- 2B (based on Qwen2-VL)
- Hardware
- from: 1 GPU
Cybersecurity2023–2024
Huawei and University of Science and Technology of China (LogAIBox project) · China
A model pretrained on a large volume of logs from different systems: it turns a log line into a vector on top of which failure and outlier detection is fine-tuned. A human makes the decision on a disputed operation; automatic blocking without review is not acceptable.
- Finding unusual entries in operation logs
- Grouping similar failure messages
- Preparing features for an in-house monitoring model
- Sizes
- about 110M (BERT-base architecture)
- Hardware
- from: Laptop
CodeOllama2023–2024
DeepSeek · China
DeepSeek's coding model family: from small autocompletion models to the large MoE V2, which matched closed models in 2024. Later, coding moved into DeepSeek's general models.
- Code autocompletion and generation
- Translating code between programming languages
- Finding bugs and explaining other people's code
- Sizes
- 1.3B – 236B-A21B
- Hardware
- from: Laptop
Fact-checking and judges2024
Flow AI · not disclosed
A small judge model: it checks an answer against your instruction and gives a score with an explanation. Fits on a modest server. The judge itself makes mistakes and does not replace manual review.
- Checking AI assistant answers against the instruction
- Bulk scoring of exported conversations
- Quality control before rolling out changes
- Sizes
- 3.8B (based on Phi-3.5-mini)
- Hardware
- from: Laptop
Forecasting2024
The Time-MoE team · not disclosed
A forecasting model with a sparse architecture: only part of the network runs at each step, so it stays fast at a small size.
- Forecasting sales and stock levels
- Forecasting load on services and staff
- Planning purchases from history
- Sizes
- 50M and 200M
- Hardware
- from: Laptop
CodeOllama2024
01.AI · China
Coding models from 01.AI at 1.5B and 9B with a 128K-token context and support for 52 programming languages. A separate line next to the text Yi models.
- Code autocompletion and generation
- Explaining and refactoring code
- A programming assistant without the cloud
- Sizes
- 1.5B – 9B
- Hardware
- from: Laptop
TextOllamaNot maintained2023–2024
Stability AI · UK
Small models from Stability AI: StableLM 2 (1.6B) knows 7 European languages, Stable Code (3B) completes code. No updates since 2024.
- A lightweight chatbot on an ordinary PC
- Code autocompletion in the editor
- A base for fine-tuning on your own task
- Sizes
- 1.6B – 12B
- Hardware
- from: Laptop
RerankersGGUFNot maintained2023–2024
BAAI (Beijing Academy of Artificial Intelligence) · China
Rerankers: they take passages found by search and reorder them by how well they actually match the question. v2-m3 is multilingual and lightweight, often paired with bge-m3.
- Refining search results before a chatbot answers
- Sorting knowledge base search results
- Selecting the most relevant clauses of contracts and policies
- Sizes
- 278M – 9B
- Hardware
- from: Laptop
CodeOllamaNot maintained2023–2024
Zhipu AI (Z.ai) and Tsinghua University · China
Coding models from the creators of GLM. CodeGeeX4-ALL-9B, based on GLM-4-9B, combines autocompletion, code chat, function calling and repository search in one model.
- Code autocompletion in the IDE
- A code chat assistant
- Answering questions about a repository
- Sizes
- 6B – 9B
- Hardware
- from: Laptop
CodeOllamaNot maintained2024
Mistral AI · France
Mistral's coding model covering 80+ programming languages. The open weights of the main version cannot be used in production without a paid license; newer Codestral versions are API-only.
- Evaluation and testing before buying a license
- Code autocompletion (with a commercial license)
- Research on coding model quality
- Sizes
- 7B – 22B
- Hardware
- from: Laptop
Fact-checking and judgesNot maintained2024
Patronus AI · USA
Checks whether a chatbot invented a fact that is not in the source documents. The license is non-commercial. The checking model itself makes mistakes and does not replace manual review on important tasks.
- Finding invented facts in AI assistant answers
- Checking that answers rest on the attached documents
- Filtering out answers before they go to a customer
- Sizes
- 8B and 70B
- Hardware
- from: 1 GPU
Text analysisRUNot maintained2020–2024
SberDevices (ai-forever) · Russia
Sber's Russian-language encoders trained on large Russian corpora. A base for classifiers, NER and semantic search in Russian.
- Classifying requests in Russian
- Extracting names, amounts and dates after fine-tuning
- Detecting review sentiment
- Sizes
- about 30M to 430M
- Hardware
- from: Laptop
Fact-checking and judgesNot maintained2024
RLHFlow · USA
An answer scorer that returns a breakdown across several attributes rather than a single overall score. The scorer itself makes mistakes and does not replace manual review on important tasks.
- Choosing the best of several candidate answers
- Preparing data for model fine-tuning
- Scoring assistant answers across several attributes
- Sizes
- 8B
- Hardware
- from: 1 GPU
TextOllamaNot maintained2023–2024
OpenChat (Tsinghua University) · China
Fine-tunes of Mistral 7B and Llama 3 8B using the C-RLFT method that caught up with ChatGPT-3.5 in 2023–2024 at just 7–8B. A lightweight general-purpose assistant for a modest server.
- Chat assistant on an inexpensive server
- Drafts of emails and replies
- Help with simple code
- Sizes
- 7B – 13B
- Hardware
- from: Laptop
Text to SQLOllamaNot maintained2023–2024
Defog · USA
One of the first open models that turn a plain-language question into an SQL query against a database. Available in Ollama, but newer competitors are already stronger.
- Answering managers' questions from the sales database without an analyst
- Drafting SQL queries for reports
- An assistant inside a BI system
- Sizes
- 7B – 70B
- Hardware
- from: Laptop
Text to SQLGGUFNot maintained2024
Chat2DB · China
A text-to-SQL model from the open Chat2DB database client: it supports different SQL dialects, with an English and Chinese model card.
- Turning a question into SQL inside a database client
- Drafting queries for different database engines
- Hints for developers working with a schema
- Sizes
- 7B
- Hardware
- from: Laptop
Fact-checking and judgesNot maintained2023–2024
KAIST and LG AI Research (prometheus-eval) · South Korea
An open judge model: it scores other models' answers against your criteria and explains the score. A replacement for paid models in the reviewer role.
- Scoring chatbot answers on your own scale
- Comparing two answer options
- Quality checks before launching an AI service
- Sizes
- 7B – 8x7B
- Hardware
- from: Laptop
CodeOllamaNot maintained2023–2024
BigCode (Hugging Face and ServiceNow) · USA / France
One of the first open coding models, trained on an open set of source code with an option to exclude your own repository. Today it is more a base for fine-tuning than a leader.
- Code autocompletion in the editor
- Fine-tuning on the company's internal code
- Generating boilerplate code and tests
- Sizes
- 1B – 15B
- Hardware
- from: Laptop
CybersecurityNot maintained2023–2024
SmartData@PoliTO, Politecnico di Torino · Italy
A pair of models from an Italian university: they read a recorded command session and mark which parts belong to reconnaissance, persistence and other attacker steps. A human makes the decision on a disputed operation; automatic blocking without review is not acceptable.
- Labelling command logs by stage of activity
- Grouping similar sessions for incident review
- Preparing features for server monitoring
- Sizes
- about 125M (based on CodeBERT)
- Hardware
- from: Laptop
TextOllamaNot maintained2023–2024
TinyLlama (SUTD researchers) · Singapore
A 1.1B model with the Llama 2 architecture, trained on 3 trillion tokens. Now behind newer small models, but still a popular base for experiments and fine-tuning.
- Simple chatbots on low-end hardware
- Experiments and team training
- A base for fine-tuning on a narrow task
- Sizes
- 1.1B
- Hardware
- from: Laptop
CybersecurityGGUFNot maintained2023–2024
ZySec AI · India
A small open assistant for security professionals: questions about standards, reviewing threats and vulnerabilities, drafting internal documents.
- Answering questions about security policies and standards
- First-pass review of threat reports
- Drafting internal protection guidelines
- Sizes
- 2.8B и 7B
- Hardware
- from: Laptop
Search and RAGRUNot maintained2022–2024
Microsoft · USA
Proven models for semantic search. The multilingual versions work well with Russian and are still a reliable base for RAG.
- Search across a knowledge base and documents
- Finding answers for a chatbot (RAG)
- Finding similar requests and duplicates
- Sizes
- 33M – 7B
- Hardware
- from: Laptop
Text to SQLGGUFNot maintained2024
ChatDB · USA
A text-to-SQL model built on DeepSeek-Coder, aimed at complex questions spanning several tables and conditions.
- Complex queries joining several tables
- Answering database questions without an analyst
- Drafting SQL for reports and exports
- Sizes
- 7B
- Hardware
- from: Laptop
CodeOllamaNot maintained2023–2024
Meta · USA
A version of Llama 2 further trained on code, with variants for Python and for chat. Outdated, but many ready-made fine-tuned versions and tools exist.
- Code autocompletion and explanation
- Generating Python scripts
- Base model for fine-tuning on your own stack
- Sizes
- 7B – 70B
- Hardware
- from: Laptop
Text to SQLOllamaNot maintained2024
MotherDuck and Numbers Station · USA
A model for turning questions into SQL, built for the embedded analytics database DuckDB. Available in Ollama, convenient for working with CSV and Parquet locally.
- Plain-language questions about CSV and Parquet exports
- DuckDB queries inside analytics scripts
- Quick analytics on a laptop without a server
- Sizes
- 7B
- Hardware
- from: Laptop
RerankersGGUFNot maintained2023–2024
University of Waterloo, Castorini group · Canada
Rerankers that are language models: they receive the whole list of retrieved passages and reorder it as a list, instead of scoring passages one by one. Heavier than ordinary rerankers.
- Reordering a long list of search results
- Selecting sources for an AI assistant answer
- Research comparisons of retrieval approaches
- Sizes
- 7B – 13B
- Hardware
- from: 1 GPU
TextOllamaNot maintained2023–2024
WizardLM (Microsoft and Peking University) · USA / China
Fine-tunes of Llama, Mistral and StarCoder using Evol-Instruct, which automatically makes instructions more complex. WizardLM-2 was released in April 2024 and removed almost immediately, so only the 2023 versions are relevant.
- Complex multi-step instructions
- Help for developers
- Solving math problems
- Sizes
- 7B – 70B
- Hardware
- from: Laptop
RerankersNot maintained2023
NetEase Youdao · China
An embedding-plus-reranker pair for knowledge bases. The card lists English, Chinese, Japanese and Korean — Russian is not among the stated languages.
- Search across a knowledge base and reference materials
- Reordering retrieved passages
- Picking answers for a support chatbot
- Sizes
- about 280M
- Hardware
- from: Laptop
Text to speechNot maintained2023
Columbia University · USA
A lightweight English speech synthesis model with natural intonation. Many other models, such as Kokoro, are built on it.
- Voicing texts in English
- A base for fine-tuning your own voice
- Voice service prototypes
- Sizes
- about 150M
- Hardware
- from: Laptop
Text to SQLNot maintained2023
RUCKBReasoning, Renmin University of China · China
An early line of open text-to-SQL models starting at 1B, including variants fine-tuned for specific database schemas.
- Turning an employee question into an SQL query
- Drafting warehouse queries for a report
- Embedding into a BI dashboard as a helper
- Sizes
- 1B – 15B
- Hardware
- from: Laptop
Text to SQLGGUFNot maintained2023
Numbers Station · USA
One of the first open text-to-SQL lines, including very small versions from 350M that run on an ordinary PC.
- Turning a question into SQL from a table description
- Hints while writing queries
- A local analyst helper with no data leaving the company
- Sizes
- 350M – 7B
- Hardware
- from: Laptop
RerankersRUNot maintained2022
UKP Lab and the Sentence Transformers community · Germany
The most downloaded open rerankers: a tiny model reads a question-passage pair and scores how well they match. The multilingual mMARCO version covers Russian.
- Reordering knowledge base search results
- Selecting passages before a chatbot answers
- Finding duplicates among tickets and product cards
- Sizes
- about 4M – 120M
- Hardware
- from: Laptop
Text analysisRUNot maintained2019–2022
Meta · USA
A classic multilingual encoder for 100 languages, including Russian. The base of many sentiment, NER and embedding models, including BGE-M3.
- Detecting review sentiment in different languages
- Extracting names and organizations after fine-tuning
- Classifying requests
- Sizes
- 270M – 10.7B
- Hardware
- from: Laptop
Text analysisRUNot maintained2021
Microsoft · USA
A time-tested encoder behind many classifiers and NER models (including GLiNER). The multilingual mDeBERTa-v3 understands Russian.
- Classifying review sentiment
- Entity extraction after fine-tuning
- Checking whether a conclusion follows from a text
- Sizes
- 70M – 435M
- Hardware
- from: Laptop
Text analysisRUNot maintained2021
David Dale (cointegrated) · Russia
A very small Russian-English BERT that runs fast on a regular CPU. Ready-made fine-tuned versions exist for sentiment, toxicity and emotions.
- Detecting review sentiment
- Filtering rude chat messages
- Fast classification of requests
- Sizes
- 12M – 29M
- Hardware
- from: Laptop
