Model comparison
Codestral vs DeepSeek-R1-Distill-Qwen-32B
DeepSeek-R1-Distill-Qwen-32B is the stronger model overall, scoring 35.0 to 30.1 on the Noometry Index.
Last verified . 2 shared benchmarks.
Mistral AI
DeepSeekSummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 1 category and DeepSeek-R1-Distill-Qwen-32B in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-R1-Distill-Qwen-32B leads 35.6 to 26.9.
Side by side
| Codestral | DeepSeek-R1-Distill-Qwen-32B | |
|---|---|---|
| Provider | Mistral AI | DeepSeek |
| Noometry Index | 30.1 | 35.0 |
| Released | 2024-05-29 | 2025-01-20 |
| Weights | Open | Open |
| Context window | 256K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.30 | — |
| Output $ / M tokens | $0.90 | — |
| Results tracked | 7 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1-Distill-Qwen-32B leads
Codestral: 26.9 (#316), DeepSeek-R1-Distill-Qwen-32B: 35.6 (#209)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| BigCodeBench Instruct | 41.8% | 43.9% |
| BigCodeBench Complete | 52.5% | 54.9% |
| Aider Polyglot | 11.1% | — |
| LiveBench Coding | — | 33.7% |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Agentic & Tool Use Not comparable
Codestral: —, DeepSeek-R1-Distill-Qwen-32B: 27.8 (#88)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| BALROG | — | 19.5% |
Reasoning Codestral leads
Codestral: 19.1 (#244), DeepSeek-R1-Distill-Qwen-32B: 17.6 (#280)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 52.3% |
| LiveBench Data Analysis | — | 45.4% |
| Epoch Capabilities Index | — | 137.43 |
| LiveBench | — | 45.5% |
Math Not comparable
Codestral: —, DeepSeek-R1-Distill-Qwen-32B: 32.9 (#183)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 55.6% |
| LiveBench Math | — | 59.4% |
Knowledge Not comparable
Codestral: —, DeepSeek-R1-Distill-Qwen-32B: 35.7 (#167)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| GPQA Diamond | — | 64.1% |
Instruction Following Not comparable
Codestral: —, DeepSeek-R1-Distill-Qwen-32B: 61.7 (#239)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| LiveBench Instruction Following | — | 55.7% |
Writing & Preference Not comparable
Codestral: —, DeepSeek-R1-Distill-Qwen-32B: 49.6 (#185)
| Benchmark | Codestral | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| LiveBench Language | — | 26.8% |
Frequently asked questions
Is Codestral better than DeepSeek-R1-Distill-Qwen-32B?
DeepSeek-R1-Distill-Qwen-32B is the stronger model overall, scoring 35.0 to 30.1 on the Noometry Index.
Is Codestral or DeepSeek-R1-Distill-Qwen-32B better for coding?
DeepSeek-R1-Distill-Qwen-32B scores higher on coding benchmarks: 35.6 versus 26.9 in the Noometry coding category.
How many benchmarks do Codestral and DeepSeek-R1-Distill-Qwen-32B share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and DeepSeek-R1-Distill-Qwen-32B has 14.