CO

Codestral

Coding model

Codestral

Release date: 2024-05-29Updated: 2024-05-30 15:12:07739
Live demoGitHubHugging FaceCompare
Parameters
22B
Context length
32K
Chinese support
Not supported
Reasoning ability

Codestral is an AI model published by MistralAI, released on 2024-05-29, for Coding model, with 22B parameters, and 32K context length, requiring about 44GB storage, under the Mistral AI Non-Production License license, with a 81.10 score on HumanEval.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Codestral

Model basics

Reasoning traces
Not supported
Thinking modes
Thinking modes not supported
Context length
32K tokens
Max output length
No data
Model type
Coding model
Modality (in / out)
No data
Release date
2024-05-29
Model file size
44GB
MoE architecture
No
Total params / Active params
22B / N/A
Knowledge cutoff
No data
Codestral

Open source & experience

Weights license
GitHub repo
GitHub link unavailable
Live demo
No live demo
Codestral

Official resources

Codestral

API details

API speed
No data
No public API pricing yet.
Codestral

Benchmark Results

Codestral currently shows benchmark results led by MBPP (12 / 28, score 78.20), HumanEval (22 / 39, score 81.10), LiveCodeBench (113 / 120, score 31.50). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking

Coding and Software Engineer

3 evaluations
Benchmark / mode
Score
Rank/total
81.10
22 / 39
78.20
12 / 28
31.50
113 / 120

Compare with other models

No curated comparisons for this model yet.

Want a custom combination? Open the compare tool

Codestral

Publisher

Codestral

Model Overview

Codestral is an AI model published by MistralAI, released on 2024-05-29, for Coding model, with 22B parameters, and 32K context length, requiring about 44GB storage, under the Mistral AI Non-Production License license, with a 81.10 score on HumanEval.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code