StarChat Models by Hugging Face H4
Last refreshed 2026-05-19. Next refresh: weekly.
Details
About
StarChat is Hugging Face H4’s series of chat models trained to act as helpful coding assistants on top of BigCode StarCoder bases. StarChat Alpha was introduced as a fine-tune of StarCoder for conversational coding help; StarChat-β is the second model in the series, a 16B GPT-like fine-tune of StarCoderPlus on an “uncensored” OpenAssistant-guanaco variant, aimed at English plus 80+ programming languages. The project frames StarChat as an open coding-chat alternative built with Transformers, with demos and checkpoints under the HuggingFaceH4 org.
Decision facts
- Best fit
- coding
- Capability starting point
- StarChat Alpha with 8K context
- Lowest tracked input
- Not tracked
- Closest related family
- StarChat2
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 8K context and 16B parameters.
Use when the workload needs 8K context and 16B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| StarChat Alpha | Use when the workload needs 8K context and 16B parameters. | 2023-05 | 8K context16B parameters | Current |
| StarChat Beta | Use when the workload needs 8K context and 16B parameters. | 2023-05 | 8K context16B parameters | Current |
Release Timeline
1 release groupSpecifications(2 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| StarChat Alpha | 2023-05 | 8K | 16B |
| StarChat Beta | 2023-05 | 8K | 16B |
