BenchMark'd
No ratings yet
Be the first to review this model
A lightweight open-source coding model from Mistral AI that runs on consumer hardware. Achieves 68% on SWE-bench Verified, rivaling much larger models. Apache 2.0 licensed for local development.
Released
December 9, 2025
Parameters
24B
Context
128K
Pricing
Free
| Benchmark | Category | Score | Performance |
|---|---|---|---|
HumanEval | coding | 81.0% | 81 |
Last updated: March 15, 2026
Benchmark scores may vary based on evaluation methodology and conditions.