Codestral and Devstral: Mistral Models for Code Work

Codestral and Devstral is a useful local-model lesson because it makes one trade-off visible: code completion, patch drafting, software-agent experiments, and comparing local code models against Qwen Coder or StarCoder2. The point is not to crown a permanent winner. The point is to learn how to match a model family to hardware, task, license, and risk.

Current source signal

Build the small version

Build a license-aware model picker for code: one field for task quality, one for runtime fit, one for allowed use.

1. 1Pick one exact model file or runtime tag from the current model card.
2. 2Run three short prompts: one easy, one task-specific, and one likely failure case.
3. 3Record load time, response speed, memory pressure, answer quality, and one surprising failure.
4. 4Write a one-paragraph recommendation: use it, do not use it, or use it only for a narrow job.

code_model_picker: fields: - model_name - local_runtime - license_allows_project_use - best_task - max_context - fallback_model rule: never choose a code model on benchmark score alone

The big idea: remember license-aware picker. Local model work is product design under constraints, not just downloading the model with the loudest leaderboard score.