Open-weight models
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | AI territory |
|---|---|
| Categories | 6 |
| Companies | 34 |
| Part of | neural networks |
| Papers, 30 days | 311 |
| Map changed | 2026-11-17 |
| Map | open the map |
Open-weight models is a territory of the AI industry on map of ai. In short: open-weight models publish trained parameters for local deployment, fine-tuning, and hosting. The map lists 34 products and companies in 6 categories.
papers311 in the last 30 days, the first 10
- Democratizing Robot-Use Agents: A Unified Harness with Open Models 2026-11-01
- Open Vision Banana: A Fully Open Recipe for Generative Visual Perception 2026-11-11
- You don't need long context 2026-11-03
- AgentBaton: Efficient Model Collaboration in Agentic Tasks 2026-10-25
- Why Pay for Long Reasoning When Rewrite Do Trick? Causal Effects of Reasoning History Compaction on Future Generation and Agent Cost 2026-11-02
- Youtu-Parsing-Omni: One Encoder, One Schema, All Modalities 2026-10-29
- Measuring AI Agents as Teammates with Collaborative Intelligence 2026-11-02
- What Effort Level Should I Choose? On the Economics of Reasoning Effort in LLM Agents 2026-10-25
- EditAgent: Rethinking Agentic Image Editing 2026-10-29
- Adversarial Speculative Reasoning Improves Long Horizon Agents 2026-11-03
The map of open-weight models has 6 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.
frontier open llms12 products and companies
Frontier open LLMs are large language models released with publicly available weights by research labs and companies, used by developers and researchers for fine-tuning, deployment, and research. Users run these models locally or on cloud infrastructure for tasks like text generation, reasoning, and specialized applications.
inference engines5 products and companies
Inference engines run open-weight language models for text generation and understanding. Machine learning engineers, application developers, and researchers use them to execute model inference locally or via APIs.
Products and companies
Papers, last 30 days
Open ground
vLLM and Ollama optimize throughput on GPUs while llama.cpp targets CPU inference, but none of them ship native hardware acceleration for Apple Neural Engine as a first-class inference path.
model hubs and hosting6 products and companies · its own article →
Platforms and services that host and distribute open-weight machine learning models, allowing researchers and developers to download, upload, and share model weights. Used by ML practitioners, researchers, and companies building on top of open models.
Products and companies
fine-tuning tools3 products and companies · its own article →
Tools that adapt open-weight models to specific tasks or datasets through parameter updates, without retraining from scratch. Used by researchers, practitioners, and companies who need to customize existing models for domain-specific applications.
Products and companies: unsloth, axolotl, torchtune.
Papers, last 30 days:- FewStepRL: Fine-tuning Few-step Diffusion Models with Reinforcement Learning 2026-11-05
- Sustainable Federated Fine-Tuning with LoRA: Environmental Impact and Performance Footprint 2026-11-12
- SLIC: Reinforcement Fine-Tuning Small LMs for Multi-Turn Analog Circuit Optimization 2026-11-10
- Measuring the Marginal Value of Verifier-Based RL Beyond Supervised Fine-Tuning 2026-11-05
- SAFTAC: Simulation-Augmented Fine-Tuning of Open-Source LLMs for Analog Circuit Design 2026-11-05
local desktop apps4 products and companies
Local desktop applications that run open-weight language models on user machines without cloud dependency. Software developers, researchers, and end users deploy these to run inference on consumer hardware.
Products and companies: lm studio, jan, gpt4all, open webui.
open image and video models4 products and companies
Open-weight image and video generation models that researchers and developers can download and run locally or self-host. Used by ML engineers, artists, and companies building on top of generative capabilities without proprietary API dependencies.
Products and companies: stable diffusion, flux, wan, hunyuan.
See also
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.