Model card
Codestral 22B.
Mistralopen-weightsUnknown paramsDense Transformer
22B parameters. Released May 2024. Supports 80+ languages. 32K context.
§ 01 · Benchmarks
Every benchmark Codestral 22B has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | CrossCodeEval | Computer Code · Code Completion | exact-match | 35.6% | #5 | 2024-05-29 | source ↗ |
| 02 | MBPP | Computer Code · Code Generation | pass@1 | 75.4% | #13 | 2024-05-29 | source ↗ |
| 03 | LiveCodeBench | Computer Code · Code Generation | pass@1 | 29.5% | #29 | 2024-03-12 | source ↗ |
| 04 | HumanEval | Computer Code · Code Generation | pass@1 | 81.1% | #35 | 2024-05-29 | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where Codestral 22B actually performs.
§ 03 · Papers
2 papers with results for Codestral 22B.
- 2024-05-29· Computer Code· 3 results
Codestral: Hello, World!
- 2024-03-12· Computer Code· 1 result
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
§ 04 · Related models
Other Mistral models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
official-blog
3
results
official-leaderboard
1
result
4 of 4 rows marked verified. · first result 2024-03-12, latest 2024-05-29.