Model comparison
Llama 4 Maverick vs Mistral Medium
Mistral Medium is the stronger model overall, scoring 36.3 to 30.9 on the Noometry Index. Llama 4 Maverick costs 9.9× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
Meta
Mistral AISummary
- They share 33 benchmarks with published results for both. Llama 4 Maverick scores higher in 1 category and Mistral Medium in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Medium leads 60.0 to 38.8.
- The biggest single-benchmark swing is WeirdML: 24.5% for Llama 4 Maverick and 43.7% for Mistral Medium.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 128K.
Side by side
| Llama 4 Maverick | Mistral Medium | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 30.9 | 36.3 |
| Released | 2025-04-05 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 4K | 262K |
| Input $ / M tokens | $0.19 | $1.50 |
| Output $ / M tokens | $0.65 | $7.50 |
| Results tracked | 54 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Medium leads
Llama 4 Maverick: 26.6 (#324), Mistral Medium: 34.2 (#243)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| SciCode | 33.1% | 40.2% |
| WeirdML | 24.5% | 43.7% |
| LMArena Coding | 1302 | 1434 |
| ALE-Bench | 172.97 | 763.98 |
| FrontierCode | — | 8% |
| SWE-bench Verified (bash only) | 21% | — |
| Aider Polyglot | 15.6% | — |
| BigCodeBench Instruct | 49.7% | — |
| BigCodeBench Complete | 61.4% | — |
Agentic & Tool Use Too close to call
Llama 4 Maverick: 28.2 (#91), Mistral Medium: 28.3 (#90)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.3% | 37.7% |
Reasoning Mistral Medium leads
Llama 4 Maverick: 10.1 (#342), Mistral Medium: 24.0 (#167)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 55.9% | 50% |
| CritPt | 0% | 0% |
| LMArena Hard Prompts | 1281 | 1426 |
| DTBench | 61.9% | 75.5% |
| LMCA | 15.9% | 26.1% |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 27.7% | — |
| NYT Connections (extended) | 8% | — |
| ARC-AGI-1 | 4.4% | — |
| EnigmaEval | 0.6% | — |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 132.2 | — |
| ForecastBench | 57.5 | — |
Math Mistral Medium leads
Llama 4 Maverick: 26.0 (#262), Mistral Medium: 28.1 (#245)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 20.6% | 32.2% |
| LMArena Math | 1299 | 1408 |
| MATH Level 5 | 73% | 81.6% |
| FrontierMath (Feb 2025 set) | 0.7% | 0.3% |
| ProofBench | — | 9% |
| Omni-MATH | 42.2% | — |
Knowledge Llama 4 Maverick leads
Llama 4 Maverick: 33.4 (#204), Mistral Medium: 25.0 (#265)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| GPQA Diamond | 67% | 59.5% |
| Humanity's Last Exam | 5.7% | 4.5% |
| Vectara Hallucination Rate | 8.2% | 22.7% |
| LMArena Expert | 1259 | 1408 |
| MMLU-Pro | 81% | — |
| Confabulations | 22.6% | — |
| GPQA (HELM) | 65% | — |
Multimodal Mistral Medium leads
Llama 4 Maverick: 31.6 (#105), Mistral Medium: 35.3 (#88)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| LMArena Vision | 1142 | 1172 |
| GeoBench | 52% | — |
| SpatialViz-Bench | 31.8% | — |
Multilingual Mistral Medium leads
Llama 4 Maverick: 42.2 (#195), Mistral Medium: 52.1 (#91)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1269 | 1408 |
| LMArena Chinese | 1277 | 1447 |
| LMArena French | 1259 | 1459 |
| LMArena German | 1291 | 1432 |
| LMArena Japanese | 1207 | 1378 |
| LMArena Korean | 1203 | 1380 |
| LMArena Russian | 1286 | 1411 |
| LMArena Spanish | 1293 | 1433 |
Instruction Following Mistral Medium leads
Llama 4 Maverick: 71.7 (#146), Mistral Medium: 73.7 (#116)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1267 | 1398 |
| IFEval | 90.8% | — |
Long Context Mistral Medium leads
Llama 4 Maverick: 31.4 (#279), Mistral Medium: 42.9 (#114)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1280 | 1406 |
| Fiction.LiveBench | 46.2% | — |
Writing & Preference Mistral Medium leads
Llama 4 Maverick: 38.8 (#252), Mistral Medium: 60.0 (#103)
| Benchmark | Llama 4 Maverick | Mistral Medium |
|---|---|---|
| LMArena Text | 1287 | 1424 |
| LMArena Creative Writing | 1267 | 1391 |
| Short-Story Creative Writing | 62% | 77.3% |
| LMArena Multi-Turn | 1289 | 1418 |
| EQ-Bench Creative Writing | 860 | — |
| WildBench | 80% | — |
Frequently asked questions
Is Llama 4 Maverick better than Mistral Medium?
Mistral Medium is the stronger model overall, scoring 36.3 to 30.9 on the Noometry Index. Llama 4 Maverick costs 9.9× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Which is cheaper, Llama 4 Maverick or Mistral Medium?
Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is Llama 4 Maverick or Mistral Medium better for coding?
Mistral Medium scores higher on coding benchmarks: 34.2 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
Mistral Medium does, with 262K tokens against 128K.
How many benchmarks do Llama 4 Maverick and Mistral Medium share?
33 benchmarks have published results for both models. Llama 4 Maverick has 54 scored results on Noometry and Mistral Medium has 36.