Stable Code Instruct 3B
Stable Code Instruct 3B adapts Stable Code 3B for software-development conversations and SQL generation. It combines public and synthetic training data with Direct Preference Optimization.
Model design and access
The official model card reports separate code and SQL evaluations. These are release-time measurements, not a promise about a particular codebase. The checkpoint should be loaded with its own tokenizer and instruction format rather than prompted as the base completion model.Stable Code Instruct 3B official model card
Practical use
Practical guidance. This model fits narrow local coding assistance where the task can be described with a modest amount of context. Validate generated SQL against the real schema and review code for assumptions about libraries and data. Its small size can help deployment, but it does not remove the need for execution checks.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | stable-code-instruct-3b |
| Context window | 16,384 tokens Published checkpoint sequence length; prompt and generated text share the window. Stable Code Instruct 3B configuration |
| Ancestry | Instruction-tuned from Stable Code 3B.Stable Code Instruct 3B official model card |
| License | Stability AI Community License.Stable Code Instruct 3B official model card |
| Focus | Code discussion, software tasks and SQL.Stable Code Instruct 3B official model card |
Related models
- Stable Code 3B
Stability AI · Stable Code
Stable Code 3B is a compact base model for code generation and completion.
Launch date not confirmedDownloadable - Stable LM 2 Chat 12B
Stability AI · Stable LM
Stable LM 2 Chat 12B is an English instruction-tuned model with documented function calling.
Launch date not confirmedResearch releaseDownloadable