LIVE
Introducing Cosmos 3 Edge20/07/26 · Hugging Face|Kimi K3, Qwen 3.8, and Anthropic's (Potential) Unravelling20/07/26 · Anthropic|At SIGGRAPH, NVIDIA Advances Graphics and Simulation With Agentic and Physical AI20/07/26 · NVIDIA|China's open-weights AI strategy is winning20/07/26|Bristol Myers Squibb Building Life Science Industry’s Most Advanced AI Factory on NVIDIA Vera Rubin20/07/26 · NVIDIA|Safety and alignment in an era of long-horizon models20/07/26 · OpenAI|Claude Fable produced a counterexample to the Jacobian Conjecture20/07/26 · Anthropic|Apply for Anthropic’s AI for Science rare disease research grants20/07/26 · Anthropic|AI advice made people less accurate but more confident – sudy19/07/26|Vincentwei1021/video-shotcraft: AI video skill for Claude Code & Codex — cinematic product videos with Remotion: 106 shot recipe cards, 161 motion previews, a production-ready template19/07/26 · Anthropic|Claude Code uses Bun written in Rust now19/07/26 · Anthropic|OpenAI reduces Codex Model Context Size from 372k to 272k19/07/26 · OpenAI|Introducing Cosmos 3 Edge20/07/26 · Hugging Face|Kimi K3, Qwen 3.8, and Anthropic's (Potential) Unravelling20/07/26 · Anthropic|At SIGGRAPH, NVIDIA Advances Graphics and Simulation With Agentic and Physical AI20/07/26 · NVIDIA|China's open-weights AI strategy is winning20/07/26|Bristol Myers Squibb Building Life Science Industry’s Most Advanced AI Factory on NVIDIA Vera Rubin20/07/26 · NVIDIA|Safety and alignment in an era of long-horizon models20/07/26 · OpenAI|Claude Fable produced a counterexample to the Jacobian Conjecture20/07/26 · Anthropic|Apply for Anthropic’s AI for Science rare disease research grants20/07/26 · Anthropic|AI advice made people less accurate but more confident – sudy19/07/26|Vincentwei1021/video-shotcraft: AI video skill for Claude Code & Codex — cinematic product videos with Remotion: 106 shot recipe cards, 161 motion previews, a production-ready template19/07/26 · Anthropic|Claude Code uses Bun written in Rust now19/07/26 · Anthropic|OpenAI reduces Codex Model Context Size from 372k to 272k19/07/26 · OpenAI|
🇫🇷 France💻 CodeReleased January 2025

Codestral

Mistral AI
Open-sourceFreeSelf-hostRGPD
Verdict

A solid pick for self-hosting and full data control.

Overview

Codestral (22B) is trained specifically on code and supports 80+ languages. Native Fill-In-the-Middle (FIM) support makes it ideal for IDE autocompletion. Non-commercial license for raw weights, but commercial API available.

Skill profile

FrenchReasoningSpeedCreativitySafety
Arena Elo1240
Verified benchmarks
SWE-Bench35%
HumanEval81.1%

What do these scores mean?

SWE-Bench35%fair

Resolving real GitHub issues under real conditions (SWE-bench Verified).

HumanEval81.1%excellent

Python code generation from specifications.

Strengths

  • Strong French quality
  • Has a free tier
  • Open-source and self-hostable
  • Competitive input pricing

Limitations

  • Light safety filters

Who is it for

A good fit if…
  • you build with a coding agent
  • you want to control cost or self-host
  • you work in French
  • you have GDPR constraints
Skip it if…
  • you want strict guardrails

Ideal use cases

  • IDE autocompletion
  • Code generation
  • Code review
  • FIM (Fill-in-the-middle)

Access & availability

Paid API per tokenFree tier availableSelf-hostable (open weights)

Key specifications

Context
256K
Input price
$0.30 $/M
Output price
$0.90 $/M
Speed
130 tok/s
Price not auditedBenchmarks not audited

Estimate your monthly cost

Per-token API pricing
$15/ month with Codestral

For the same usage

Indicative estimate based on standard API rates (excluding caching, batch and volume discounts). Always check the official pricing before committing.

Privacy

GDPR-compliantHosting : EU/selfAnonymizable data

Run it locally

Desktop with solid GPU / 32GB+ Mac13k1k
QuantizationDiskRAM / VRAMTypical hardware
Q4 · recommended12.7 GB16 GB32GB PC / 24GB Mac / RTX 4070 Ti+
Q8 · balanced23.8 GB28 GB64GB Mac / dual 24GB GPUs
FP16 · max quality44.4 GB51 GB64GB Mac / dual 24GB GPUs

Estimates for a moderate context. Long contexts need more RAM (KV cache).

🤗 Hugging Face
ollama run codestral
Advanced data · for experts
Architecture, modalities, detailed cost, full benchmarks
Architecture
Dense
Size
22B
Cutoff date
Jul 2024
Inputs
text
Outputs
code
License
mistral-non-commercial
Hosting
EU/self
Hallucination score
3/5
Estimated cost (API)
Typical exchange (~3k in / 1k out)
≈ $0.0018
1M in + 1M out
≈ $1.20
All benchmarks
Arena Elo1240
MMLU
GPQA
HumanEval81.1%
SWE-Bench35.0%
MATH

Similar models