Model comparison
Llama 4 Maverick vs Mistral Large 3
Mistral Large 3 is the stronger model overall, scoring 39.1 to 30.9 on the Noometry Index.
Last verified . 22 shared benchmarks.
Meta
Mistral AISummary
- They share 22 benchmarks with published results for both. Llama 4 Maverick scores higher in 0 categories and Mistral Large 3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Large 3 leads 60.0 to 38.8.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 8.2% for Llama 4 Maverick and 14.5% for Mistral Large 3.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $0.25 / $0.75 for Mistral Large 3.
- Mistral Large 3 accepts more context: 262K tokens versus 128K.
Side by side
| Llama 4 Maverick | Mistral Large 3 | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 30.9 | 39.1 |
| Released | 2025-04-05 | 2025-12-02 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 4K | 8K |
| Input $ / M tokens | $0.19 | $0.25 |
| Output $ / M tokens | $0.65 | $0.75 |
| Results tracked | 54 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Large 3 leads
Llama 4 Maverick: 26.6 (#324), Mistral Large 3: 34.4 (#237)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Coding | 1302 | 1448 |
| SWE-bench Verified (bash only) | 21% | — |
| Aider Polyglot | 15.6% | — |
| LMArena WebDev | — | 1230 |
| SciCode | 33.1% | — |
| WeirdML | 24.5% | — |
| BigCodeBench Instruct | 49.7% | — |
| BigCodeBench Complete | 61.4% | — |
| ALE-Bench | 172.97 | — |
Agentic & Tool Use Not comparable
Llama 4 Maverick: 28.2 (#91), Mistral Large 3: —
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.3% | — |
Reasoning Mistral Large 3 leads
Llama 4 Maverick: 10.1 (#342), Mistral Large 3: 15.2 (#319)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| Kagi LLM Benchmark | 55.9% | 50.9% |
| NYT Connections (extended) | 8% | 7.5% |
| LMArena Hard Prompts | 1281 | 1429 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 27.7% | — |
| ARC-AGI-1 | 4.4% | — |
| CritPt | 0% | — |
| EnigmaEval | 0.6% | — |
| Thematic Generalization | — | 23% |
| DTBench | 61.9% | — |
| LMCA | 15.9% | — |
| Epoch Capabilities Index | 132.2 | — |
| ForecastBench | 57.5 | — |
Math Mistral Large 3 leads
Llama 4 Maverick: 26.0 (#262), Mistral Large 3: 38.7 (#129)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Math | 1299 | 1414 |
| OTIS Mock AIME 2024-2025 | 20.6% | — |
| Omni-MATH | 42.2% | — |
| MATH Level 5 | 73% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |
Knowledge Mistral Large 3 leads
Llama 4 Maverick: 33.4 (#204), Mistral Large 3: 36.0 (#177)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| Vectara Hallucination Rate | 8.2% | 14.5% |
| LMArena Expert | 1259 | 1421 |
| GPQA Diamond | 67% | — |
| Humanity's Last Exam | 5.7% | — |
| MMLU-Pro | 81% | — |
| Confabulations | 22.6% | — |
| GPQA (HELM) | 65% | — |
Multimodal Mistral Large 3 leads
Llama 4 Maverick: 31.6 (#105), Mistral Large 3: 38.2 (#66)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Vision | 1142 | 1221 |
| GeoBench | 52% | — |
| SpatialViz-Bench | 31.8% | — |
Multilingual Mistral Large 3 leads
Llama 4 Maverick: 42.2 (#195), Mistral Large 3: 52.5 (#84)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Non-English | 1269 | 1413 |
| LMArena Chinese | 1277 | 1447 |
| LMArena French | 1259 | 1455 |
| LMArena German | 1291 | 1437 |
| LMArena Japanese | 1207 | 1394 |
| LMArena Korean | 1203 | 1384 |
| LMArena Russian | 1286 | 1411 |
| LMArena Spanish | 1293 | 1440 |
Instruction Following Mistral Large 3 leads
Llama 4 Maverick: 71.7 (#146), Mistral Large 3: 74.0 (#108)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Instruction Following | 1267 | 1403 |
| IFEval | 90.8% | — |
Long Context Mistral Large 3 leads
Llama 4 Maverick: 31.4 (#279), Mistral Large 3: 43.1 (#105)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Longer Query | 1280 | 1413 |
| Fiction.LiveBench | 46.2% | — |
Writing & Preference Mistral Large 3 leads
Llama 4 Maverick: 38.8 (#252), Mistral Large 3: 60.0 (#101)
| Benchmark | Llama 4 Maverick | Mistral Large 3 |
|---|---|---|
| LMArena Text | 1287 | 1428 |
| LMArena Creative Writing | 1267 | 1386 |
| EQ-Bench Creative Writing | 860 | 1412 |
| LMArena Multi-Turn | 1289 | 1429 |
| Short-Story Creative Writing | 62% | — |
| WildBench | 80% | — |
Frequently asked questions
Is Llama 4 Maverick better than Mistral Large 3?
Mistral Large 3 is the stronger model overall, scoring 39.1 to 30.9 on the Noometry Index.
Which is cheaper, Llama 4 Maverick or Mistral Large 3?
Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; Mistral Large 3 lists at $0.25 and $0.75.
Is Llama 4 Maverick or Mistral Large 3 better for coding?
Mistral Large 3 scores higher on coding benchmarks: 34.4 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
Mistral Large 3 does, with 262K tokens against 128K.
How many benchmarks do Llama 4 Maverick and Mistral Large 3 share?
22 benchmarks have published results for both models. Llama 4 Maverick has 54 scored results on Noometry and Mistral Large 3 has 24.