Model comparison
GPT-6 Sol vs Llama 4 Maverick
GPT-6 Sol is the stronger model overall, scoring 61.8 to 30.9 on the Noometry Index. Llama 4 Maverick costs 13× less per token, which makes it the better buy when GPT-6 Sol's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
MetaSummary
- They share 31 benchmarks with published results for both. GPT-6 Sol scores higher in 10 categories and Llama 4 Maverick in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Sol leads 74.0 to 10.1.
- The biggest single-benchmark swing is ARC-AGI-1: 95.5% for GPT-6 Sol and 4.4% for Llama 4 Maverick.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $2 / $10 for GPT-6 Sol.
- GPT-6 Sol accepts more context: 1.05M tokens versus 128K.
- Llama 4 Maverick has downloadable open weights; the other is API-only.
Side by side
| GPT-6 Sol | Llama 4 Maverick | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 61.8 | 30.9 |
| Released | 2026-09-22 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 128K | 4K |
| Input $ / M tokens | $2 | $0.19 |
| Output $ / M tokens | $10 | $0.65 |
| Results tracked | 45 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Sol leads
GPT-6 Sol: 60.1 (#11), Llama 4 Maverick: 26.6 (#324)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| SciCode | 57.6% | 33.1% |
| LMArena Coding | 1447 | 1302 |
| ALE-Bench | 2,462 | 172.97 |
| DeepSWE | 68.8% | — |
| FrontierCode | 49.3% | — |
| SWE-bench Verified (bash only) | — | 21% |
| Aider Polyglot | — | 15.6% |
| LMArena WebDev | 1688 | — |
| WeirdML | — | 24.5% |
| BigCodeBench Instruct | — | 49.7% |
| BigCodeBench Complete | — | 61.4% |
Agentic & Tool Use GPT-6 Sol leads
GPT-6 Sol: 37.2 (#36), Llama 4 Maverick: 28.2 (#91)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| APEX-Agents | 54.3% | — |
| Berkeley Function Calling Leaderboard | — | 37.3% |
| GDP.pdf | 26.4% | — |
| Vending-Bench 2 | 14,428 | — |
Reasoning GPT-6 Sol leads
GPT-6 Sol: 74.0 (#9), Llama 4 Maverick: 10.1 (#342)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| ARC-AGI-2 | 89.6% | 0% |
| NYT Connections (extended) | 90.1% | 8% |
| ARC-AGI-1 | 95.5% | 4.4% |
| CritPt | 30.9% | 0% |
| LMArena Hard Prompts | 1418 | 1281 |
| DTBench | 97.3% | 61.9% |
| LMCA | 59.1% | 15.9% |
| Epoch Capabilities Index | 162.72 | 132.2 |
| SimpleBench | — | 27.7% |
| Kagi LLM Benchmark | — | 55.9% |
| EnigmaEval | — | 0.6% |
| EBR-Bench | 53.3% | — |
| Mystery Game Puzzles | 56% | — |
| ForecastBench | — | 57.5 |
Math GPT-6 Sol leads
GPT-6 Sol: 87.2 (#7), Llama 4 Maverick: 26.0 (#262)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 100% | 20.6% |
| LMArena Math | 1402 | 1299 |
| FrontierMath (Tiers 1-3) | 89.8% | — |
| FrontierMath Tier 4 | 90% | — |
| ProofBench | 83% | — |
| Omni-MATH | — | 42.2% |
| MATH Level 5 | — | 73% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge GPT-6 Sol leads
GPT-6 Sol: 64.8 (#15), Llama 4 Maverick: 33.4 (#204)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 94.3% | 67% |
| Vectara Hallucination Rate | 6.5% | 8.2% |
| LMArena Expert | 1439 | 1259 |
| Humanity's Last Exam | — | 5.7% |
| SimpleQA Verified | 60.7% | — |
| MMLU-Pro | — | 81% |
| Confabulations | — | 22.6% |
| GPQA (HELM) | — | 65% |
Multimodal GPT-6 Sol leads
GPT-6 Sol: 47.6 (#10), Llama 4 Maverick: 31.6 (#105)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | 1245 | 1142 |
| GeoBench | — | 52% |
| Blueprint-Bench 2 | 36.9% | — |
| Furniture Assembly | 58.3% | — |
| SpatialViz-Bench | — | 31.8% |
Multilingual GPT-6 Sol leads
GPT-6 Sol: 50.5 (#118), Llama 4 Maverick: 42.2 (#195)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1385 | 1269 |
| LMArena Chinese | 1405 | 1277 |
| LMArena French | 1410 | 1259 |
| LMArena German | 1390 | 1291 |
| LMArena Japanese | 1385 | 1207 |
| LMArena Korean | 1341 | 1203 |
| LMArena Russian | 1401 | 1286 |
| LMArena Spanish | 1384 | 1293 |
Instruction Following GPT-6 Sol leads
GPT-6 Sol: 74.5 (#94), Llama 4 Maverick: 71.7 (#146)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| LMArena Instruction Following | 1412 | 1267 |
| IFEval | — | 90.8% |
Long Context GPT-6 Sol leads
GPT-6 Sol: 43.1 (#108), Llama 4 Maverick: 31.4 (#279)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| LMArena Longer Query | 1411 | 1280 |
| Fiction.LiveBench | — | 46.2% |
Writing & Preference GPT-6 Sol leads
GPT-6 Sol: 71.9 (#18), Llama 4 Maverick: 38.8 (#252)
| Benchmark | GPT-6 Sol | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1395 | 1287 |
| LMArena Creative Writing | 1378 | 1267 |
| EQ-Bench Creative Writing | 2125 | 860 |
| LMArena Multi-Turn | 1412 | 1289 |
| Short-Story Creative Writing | — | 62% |
| WildBench | — | 80% |
Frequently asked questions
Is GPT-6 Sol better than Llama 4 Maverick?
GPT-6 Sol is the stronger model overall, scoring 61.8 to 30.9 on the Noometry Index. Llama 4 Maverick costs 13× less per token, which makes it the better buy when GPT-6 Sol's lead doesn't matter for your workload.
Which is cheaper, GPT-6 Sol 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; GPT-6 Sol lists at $2 and $10.
Is GPT-6 Sol or Llama 4 Maverick better for coding?
GPT-6 Sol scores higher on coding benchmarks: 60.1 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Sol does, with 1.05M tokens against 128K.
How many benchmarks do GPT-6 Sol and Llama 4 Maverick share?
31 benchmarks have published results for both models. GPT-6 Sol has 45 scored results on Noometry and Llama 4 Maverick has 54.