Model comparison
Claude Opus 4 vs Llama 4 Maverick
Claude Opus 4 is the stronger model overall, scoring 43.1 to 30.9 on the Noometry Index. Llama 4 Maverick costs 99× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Last verified . 47 shared benchmarks.
MetaSummary
- They share 47 benchmarks with published results for both. Claude Opus 4 scores higher in 9 categories and Llama 4 Maverick in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Claude Opus 4 leads 61.2 to 38.8.
- The biggest single-benchmark swing is Aider Polyglot: 72% for Claude Opus 4 and 15.6% for Llama 4 Maverick.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $15 / $75 for Claude Opus 4.
- Claude Opus 4 accepts more context: 200K tokens versus 128K.
- Llama 4 Maverick has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4 | Llama 4 Maverick | |
|---|---|---|
| Provider | Anthropic | Meta |
| Noometry Index | 43.1 | 30.9 |
| Released | 2025-05-22 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 200K | 128K |
| Max output | 32K | 4K |
| Input $ / M tokens | $15 | $0.19 |
| Output $ / M tokens | $75 | $0.65 |
| Results tracked | 56 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4 leads
Claude Opus 4: 47.2 (#62), Llama 4 Maverick: 26.6 (#324)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| SWE-bench Verified (bash only) | 67.6% | 21% |
| Aider Polyglot | 72% | 15.6% |
| WeirdML | 43.7% | 24.5% |
| LMArena Coding | 1442 | 1302 |
| SWE-bench Verified | 70.7% | — |
| SciCode | — | 33.1% |
| GSO | 6.9% | — |
| BigCodeBench Instruct | — | 49.7% |
| BigCodeBench Complete | — | 61.4% |
| ALE-Bench | — | 172.97 |
| AlgoTune | 1.33 | — |
Agentic & Tool Use Claude Opus 4 leads
Claude Opus 4: 34.8 (#42), Llama 4 Maverick: 28.2 (#91)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.3% |
| Cybench | 38% | — |
| DeepResearch Bench | 46.8% | — |
| LMArena Search | 1127 | — |
| METR Time Horizons | 63.9% | — |
Reasoning Claude Opus 4 leads
Claude Opus 4: 27.3 (#121), Llama 4 Maverick: 10.1 (#342)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| ARC-AGI-2 | 8.6% | 0% |
| SimpleBench | 58.8% | 27.7% |
| Kagi LLM Benchmark | 74.3% | 55.9% |
| ARC-AGI-1 | 35.7% | 4.4% |
| CritPt | 0.3% | 0% |
| EnigmaEval | 5.6% | 0.6% |
| LMArena Hard Prompts | 1399 | 1281 |
| DTBench | 81.6% | 61.9% |
| LMCA | 37.4% | 15.9% |
| Epoch Capabilities Index | 142.67 | 132.2 |
| ForecastBench | 61.1 | 57.5 |
| NYT Connections (extended) | — | 8% |
Math Claude Opus 4 leads
Claude Opus 4: 42.0 (#86), Llama 4 Maverick: 26.0 (#262)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 20.6% |
| Omni-MATH | 61.6% | 42.2% |
| LMArena Math | 1390 | 1299 |
| MATH Level 5 | 85% | 73% |
| FrontierMath (Feb 2025 set) | 4.5% | 0.7% |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge Claude Opus 4 leads
Claude Opus 4: 44.0 (#88), Llama 4 Maverick: 33.4 (#204)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 76.3% | 67% |
| Humanity's Last Exam | 10.7% | 5.7% |
| MMLU-Pro | 87.5% | 81% |
| Confabulations | 15.9% | 22.6% |
| Vectara Hallucination Rate | 12% | 8.2% |
| GPQA (HELM) | 70.8% | 65% |
| LMArena Expert | 1386 | 1259 |
Multimodal Too close to call
Claude Opus 4: 31.5 (#106), Llama 4 Maverick: 31.6 (#105)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | 1192 | 1142 |
| GeoBench | 49% | 52% |
| VPCT | 38% | — |
| SpatialViz-Bench | — | 31.8% |
Multilingual Claude Opus 4 leads
Claude Opus 4: 48.8 (#138), Llama 4 Maverick: 42.2 (#195)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1362 | 1269 |
| LMArena Chinese | 1386 | 1277 |
| LMArena French | 1372 | 1259 |
| LMArena German | 1391 | 1291 |
| LMArena Japanese | 1331 | 1207 |
| LMArena Korean | 1321 | 1203 |
| LMArena Russian | 1392 | 1286 |
| LMArena Spanish | 1389 | 1293 |
Instruction Following Claude Opus 4 leads
Claude Opus 4: 77.1 (#28), Llama 4 Maverick: 71.7 (#146)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| IFEval | 91.8% | 90.8% |
| LMArena Instruction Following | 1406 | 1267 |
Long Context Claude Opus 4 leads
Claude Opus 4: 39.6 (#172), Llama 4 Maverick: 31.4 (#279)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| Fiction.LiveBench | 61.1% | 46.2% |
| LMArena Longer Query | 1422 | 1280 |
Writing & Preference Claude Opus 4 leads
Claude Opus 4: 61.2 (#89), Llama 4 Maverick: 38.8 (#252)
| Benchmark | Claude Opus 4 | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1377 | 1287 |
| LMArena Creative Writing | 1387 | 1267 |
| Short-Story Creative Writing | 83.6% | 62% |
| EQ-Bench Creative Writing | 1580 | 860 |
| WildBench | 85.2% | 80% |
| LMArena Multi-Turn | 1396 | 1289 |
Frequently asked questions
Is Claude Opus 4 better than Llama 4 Maverick?
Claude Opus 4 is the stronger model overall, scoring 43.1 to 30.9 on the Noometry Index. Llama 4 Maverick costs 99× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4 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; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or Llama 4 Maverick better for coding?
Claude Opus 4 scores higher on coding benchmarks: 47.2 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
Claude Opus 4 does, with 200K tokens against 128K.
How many benchmarks do Claude Opus 4 and Llama 4 Maverick share?
47 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and Llama 4 Maverick has 54.