Model comparison
GPT-4.1 vs Llama 4 Maverick
GPT-4.1 is the stronger model overall, scoring 35.9 to 30.9 on the Noometry Index. Llama 4 Maverick costs 12× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Last verified . 46 shared benchmarks.
MetaSummary
- They share 46 benchmarks with published results for both. GPT-4.1 scores higher in 8 categories and Llama 4 Maverick in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-4.1 leads 57.6 to 38.8.
- The biggest single-benchmark swing is Aider Polyglot: 52.4% for GPT-4.1 and 15.6% for Llama 4 Maverick.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 128K.
- Llama 4 Maverick has downloadable open weights; the other is API-only.
Side by side
| GPT-4.1 | Llama 4 Maverick | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 35.9 | 30.9 |
| Released | 2025-04-14 | 2025-04-05 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 33K | 4K |
| Input $ / M tokens | $2 | $0.19 |
| Output $ / M tokens | $8 | $0.65 |
| Results tracked | 52 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4.1 leads
GPT-4.1: 34.4 (#238), Llama 4 Maverick: 26.6 (#324)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| SWE-bench Verified (bash only) | 39.6% | 21% |
| Aider Polyglot | 52.4% | 15.6% |
| WeirdML | 39% | 24.5% |
| LMArena Coding | 1391 | 1302 |
| ALE-Bench | 558.1 | 172.97 |
| SWE-bench Verified | 48.5% | — |
| SciCode | — | 33.1% |
| BigCodeBench Instruct | — | 49.7% |
| BigCodeBench Complete | — | 61.4% |
| CadEval | 42% | — |
Agentic & Tool Use GPT-4.1 leads
GPT-4.1: 34.7 (#43), Llama 4 Maverick: 28.2 (#91)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | 54% | 37.3% |
Reasoning GPT-4.1 leads
GPT-4.1: 11.7 (#339), Llama 4 Maverick: 10.1 (#342)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| ARC-AGI-2 | 0.4% | 0% |
| SimpleBench | 27% | 27.7% |
| Kagi LLM Benchmark | 52.3% | 55.9% |
| ARC-AGI-1 | 5.5% | 4.4% |
| EnigmaEval | 2.2% | 0.6% |
| LMArena Hard Prompts | 1384 | 1281 |
| DTBench | 68.3% | 61.9% |
| LMCA | 25.6% | 15.9% |
| Epoch Capabilities Index | 136.78 | 132.2 |
| ForecastBench | 61.5 | 57.5 |
| NYT Connections (extended) | — | 8% |
| CritPt | — | 0% |
| Chess Puzzles | 6% | — |
Math Llama 4 Maverick leads
GPT-4.1: 22.3 (#280), Llama 4 Maverick: 26.0 (#262)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 38.3% | 20.6% |
| Omni-MATH | 47.1% | 42.2% |
| LMArena Math | 1370 | 1299 |
| MATH Level 5 | 83% | 73% |
| FrontierMath (Feb 2025 set) | 5.5% | 0.7% |
| FrontierMath (Tiers 1-3) | 6% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge GPT-4.1 leads
GPT-4.1: 37.1 (#160), Llama 4 Maverick: 33.4 (#204)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 66.9% | 67% |
| Humanity's Last Exam | 5.4% | 5.7% |
| MMLU-Pro | 81.1% | 81% |
| Vectara Hallucination Rate | 5.6% | 8.2% |
| GPQA (HELM) | 65.9% | 65% |
| LMArena Expert | 1364 | 1259 |
| SimpleQA Verified | 31.1% | — |
| Confabulations | — | 22.6% |
Multimodal GPT-4.1 leads
GPT-4.1: 38.2 (#67), Llama 4 Maverick: 31.6 (#105)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | 1211 | 1142 |
| GeoBench | 72% | 52% |
| SpatialViz-Bench | — | 31.8% |
Multilingual GPT-4.1 leads
GPT-4.1: 49.4 (#133), Llama 4 Maverick: 42.2 (#195)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1370 | 1269 |
| LMArena Chinese | 1382 | 1277 |
| LMArena French | 1382 | 1259 |
| LMArena German | 1381 | 1291 |
| LMArena Japanese | 1319 | 1207 |
| LMArena Korean | 1339 | 1203 |
| LMArena Russian | 1377 | 1286 |
| LMArena Spanish | 1376 | 1293 |
Instruction Following Too close to call
GPT-4.1: 71.3 (#153), Llama 4 Maverick: 71.7 (#146)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| IFEval | 83.8% | 90.8% |
| LMArena Instruction Following | 1367 | 1267 |
Long Context GPT-4.1 leads
GPT-4.1: 40.0 (#163), Llama 4 Maverick: 31.4 (#279)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| Fiction.LiveBench | 63.9% | 46.2% |
| LMArena Longer Query | 1385 | 1280 |
Writing & Preference GPT-4.1 leads
GPT-4.1: 57.6 (#125), Llama 4 Maverick: 38.8 (#252)
| Benchmark | GPT-4.1 | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1383 | 1287 |
| LMArena Creative Writing | 1363 | 1267 |
| EQ-Bench Creative Writing | 1420 | 860 |
| WildBench | 85.4% | 80% |
| LMArena Multi-Turn | 1398 | 1289 |
| Short-Story Creative Writing | — | 62% |
Frequently asked questions
Is GPT-4.1 better than Llama 4 Maverick?
GPT-4.1 is the stronger model overall, scoring 35.9 to 30.9 on the Noometry Index. Llama 4 Maverick costs 12× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Which is cheaper, GPT-4.1 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-4.1 lists at $2 and $8.
Is GPT-4.1 or Llama 4 Maverick better for coding?
GPT-4.1 scores higher on coding benchmarks: 34.4 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 128K.
How many benchmarks do GPT-4.1 and Llama 4 Maverick share?
46 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and Llama 4 Maverick has 54.