Model comparison
Codestral vs GPT-4.1 nano
Codestral is the stronger model overall, scoring 30.6 to 27.9 on the Noometry Index. GPT-4.1 nano costs 2.6× less per token, which makes it the better buy when Codestral's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Mistral AISummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 2 categories and GPT-4.1 nano in 0 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Codestral leads 19.8 to 8.5.
- GPT-4.1 nano is cheaper at $0.10 / $0.40 per million input/output tokens, against $0.30 / $0.90 for Codestral.
- GPT-4.1 nano accepts more context: 1.05M tokens versus 256K.
Side by side
| Codestral | GPT-4.1 nano | |
|---|---|---|
| Provider | Mistral AI | OpenAI |
| Noometry Index | 30.6 | 27.9 |
| Released | 2024-05-29 | 2025-04-14 |
| Weights | Proprietary | Proprietary |
| Context window | 256K | 1.05M |
| Max output | 8K | 33K |
| Input $ / M tokens | $0.30 | $0.10 |
| Output $ / M tokens | $0.90 | $0.40 |
| Results tracked | 7 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Codestral leads
Codestral: 27.3 (#321), GPT-4.1 nano: 24.1 (#330)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| Aider Polyglot | 11.1% | 8.9% |
| SciCode | — | 25.9% |
| WeirdML | — | 19% |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1306 |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Agentic & Tool Use Not comparable
Codestral: —, GPT-4.1 nano: 26.5 (#104)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 33% |
Reasoning Codestral leads
Codestral: 19.8 (#251), GPT-4.1 nano: 8.5 (#349)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | 33.3% |
| ARC-AGI-2 | — | 0% |
| ARC-AGI-1 | — | 0% |
| CritPt | — | 0% |
| LMArena Hard Prompts | — | 1286 |
| DTBench | — | 52.5% |
| LMCA | — | 5.5% |
| Epoch Capabilities Index | — | 129.62 |
Math Not comparable
Codestral: —, GPT-4.1 nano: 26.9 (#252)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 28.9% |
| Omni-MATH | — | 36.7% |
| LMArena Math | — | 1274 |
| MATH Level 5 | — | 70% |
| FrontierMath (Feb 2025 set) | — | 1% |
Knowledge Not comparable
Codestral: —, GPT-4.1 nano: 21.8 (#273)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| GPQA Diamond | — | 48.9% |
| SimpleQA Verified | — | 6% |
| MMLU-Pro | — | 55% |
| GPQA (HELM) | — | 50.7% |
| LMArena Expert | — | 1272 |
Multimodal Not comparable
Codestral: —, GPT-4.1 nano: 29.2 (#113)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| LMArena Vision | — | 1063 |
Multilingual Not comparable
Codestral: —, GPT-4.1 nano: 41.6 (#205)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| LMArena Non-English | — | 1260 |
| LMArena Chinese | — | 1270 |
| LMArena German | — | 1288 |
| LMArena Japanese | — | 1198 |
| LMArena Russian | — | 1261 |
Instruction Following Not comparable
Codestral: —, GPT-4.1 nano: 67.8 (#193)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| IFEval | — | 84.3% |
| LMArena Instruction Following | — | 1267 |
Long Context Not comparable
Codestral: —, GPT-4.1 nano: 23.7 (#296)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| Fiction.LiveBench | — | 25% |
| LMArena Longer Query | — | 1283 |
Writing & Preference Not comparable
Codestral: —, GPT-4.1 nano: 40.5 (#243)
| Benchmark | Codestral | GPT-4.1 nano |
|---|---|---|
| LMArena Text | — | 1285 |
| LMArena Creative Writing | — | 1260 |
| EQ-Bench Creative Writing | — | 946 |
| WildBench | — | 81.2% |
| LMArena Multi-Turn | — | 1277 |
Frequently asked questions
Is Codestral better than GPT-4.1 nano?
Codestral is the stronger model overall, scoring 30.6 to 27.9 on the Noometry Index. GPT-4.1 nano costs 2.6× less per token, which makes it the better buy when Codestral's lead doesn't matter for your workload.
Which is cheaper, Codestral or GPT-4.1 nano?
GPT-4.1 nano is cheaper. It lists at $0.10 per million input tokens and $0.40 per million output tokens; Codestral lists at $0.30 and $0.90.
Is Codestral or GPT-4.1 nano better for coding?
Codestral scores higher on coding benchmarks: 27.3 versus 24.1 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 nano does, with 1.05M tokens against 256K.
How many benchmarks do Codestral and GPT-4.1 nano share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and GPT-4.1 nano has 38.