Open-Assistant Models by OpenAssistant
Last refreshed 2026-05-19. Next refresh: weekly.
Details
About
OpenAssistant, developed by LAION, is a family of open-source large language models (LLMs) designed to serve as accessible, powerful chat-based assistants akin to models like ChatGPT and GPT-4. The initiative focused on broad accessibility, including commercial use, setting it apart from other open-source LLMs with limited licenses. These models were trained on a vast dataset gathered through a community effort with over 13,000 volunteers contributing more than 600,000 interactions. The training regimen included supervised fine-tuning and reinforcement learning from human feedback, enhancing the models' ability to follow instructions and be helpful. Although the project has ended, the models and datasets remain available for public use. Some models, however, might be based on more restricted models like LLaMA 67810.
Decision facts
- Best fit
- General model comparison
- Capability starting point
- Open-Assistant SFT-1 12B with 2k context
- Lowest tracked input
- Not tracked
- Closest related family
- Claude 3
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 2k context and 12B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Open-Assistant SFT-1 12B | Use when the workload needs 2k context and 12B parameters. | 2023-04 | 2k context12B parameters | Current |
Release Timeline
1 release groupSpecifications(1 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| Open-Assistant SFT-1 12B | 2023-04 | 2k | 12B |
