Model comparison
Codestral vs Llama 4 Maverick
Codestral and Llama 4 Maverick score almost the same on the Noometry Index (30.6 vs 30.9), so choose on price, context window or the category you care about most.
Last verified . 5 shared benchmarks.
Mistral AI
MetaSummary
- They share 5 benchmarks with published results for both. Codestral scores higher in 2 categories and Llama 4 Maverick in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in reasoning, where Codestral leads 19.8 to 10.1.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 32.5% for Codestral and 55.9% for Llama 4 Maverick.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $0.30 / $0.90 for Codestral.
- Codestral accepts more context: 256K tokens versus 128K.
- Llama 4 Maverick has downloadable open weights; the other is API-only.
Side by side
| Codestral | Llama 4 Maverick | |
|---|---|---|
| Provider | Mistral AI | Meta |
| Noometry Index | 30.6 | 30.9 |
| Released | 2024-05-29 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 256K | 128K |
| Max output | 8K | 4K |
| Input $ / M tokens | $0.30 | $0.19 |
| Output $ / M tokens | $0.90 | $0.65 |
| Results tracked | 7 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Codestral: 27.3 (#321), Llama 4 Maverick: 26.6 (#324)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| Aider Polyglot | 11.1% | 15.6% |
| BigCodeBench Instruct | 41.8% | 49.7% |
| BigCodeBench Complete | 52.5% | 61.4% |
| ALE-Bench | 137.78 | 172.97 |
| SWE-bench Verified (bash only) | — | 21% |
| SciCode | — | 33.1% |
| WeirdML | — | 24.5% |
| LMArena Coding | — | 1302 |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Agentic & Tool Use Not comparable
Codestral: —, Llama 4 Maverick: 28.2 (#91)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.3% |
Reasoning Codestral leads
Codestral: 19.8 (#251), Llama 4 Maverick: 10.1 (#342)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | 55.9% |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 27.7% |
| NYT Connections (extended) | — | 8% |
| ARC-AGI-1 | — | 4.4% |
| CritPt | — | 0% |
| EnigmaEval | — | 0.6% |
| LMArena Hard Prompts | — | 1281 |
| DTBench | — | 61.9% |
| LMCA | — | 15.9% |
| Epoch Capabilities Index | — | 132.2 |
| ForecastBench | — | 57.5 |
Math Not comparable
Codestral: —, Llama 4 Maverick: 26.0 (#262)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 20.6% |
| Omni-MATH | — | 42.2% |
| LMArena Math | — | 1299 |
| MATH Level 5 | — | 73% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Not comparable
Codestral: —, Llama 4 Maverick: 33.4 (#204)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | — | 67% |
| Humanity's Last Exam | — | 5.7% |
| MMLU-Pro | — | 81% |
| Confabulations | — | 22.6% |
| Vectara Hallucination Rate | — | 8.2% |
| GPQA (HELM) | — | 65% |
| LMArena Expert | — | 1259 |
Multimodal Not comparable
Codestral: —, Llama 4 Maverick: 31.6 (#105)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | — | 1142 |
| GeoBench | — | 52% |
| SpatialViz-Bench | — | 31.8% |
Multilingual Not comparable
Codestral: —, Llama 4 Maverick: 42.2 (#195)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | — | 1269 |
| LMArena Chinese | — | 1277 |
| LMArena French | — | 1259 |
| LMArena German | — | 1291 |
| LMArena Japanese | — | 1207 |
| LMArena Korean | — | 1203 |
| LMArena Russian | — | 1286 |
| LMArena Spanish | — | 1293 |
Instruction Following Not comparable
Codestral: —, Llama 4 Maverick: 71.7 (#146)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| IFEval | — | 90.8% |
| LMArena Instruction Following | — | 1267 |
Long Context Not comparable
Codestral: —, Llama 4 Maverick: 31.4 (#279)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| Fiction.LiveBench | — | 46.2% |
| LMArena Longer Query | — | 1280 |
Writing & Preference Not comparable
Codestral: —, Llama 4 Maverick: 38.8 (#252)
| Benchmark | Codestral | Llama 4 Maverick |
|---|---|---|
| LMArena Text | — | 1287 |
| LMArena Creative Writing | — | 1267 |
| Short-Story Creative Writing | — | 62% |
| EQ-Bench Creative Writing | — | 860 |
| WildBench | — | 80% |
| LMArena Multi-Turn | — | 1289 |
Frequently asked questions
Is Codestral better than Llama 4 Maverick?
Codestral and Llama 4 Maverick score almost the same on the Noometry Index (30.6 vs 30.9), so choose on price, context window or the category you care about most.
Which is cheaper, Codestral or Llama 4 Maverick?
Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; Codestral lists at $0.30 and $0.90.
Is Codestral or Llama 4 Maverick better for coding?
They score almost the same on coding (27.3 vs 26.6); test both on your own repository before choosing.
Which has the bigger context window?
Codestral does, with 256K tokens against 128K.
How many benchmarks do Codestral and Llama 4 Maverick share?
5 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Llama 4 Maverick has 54.