Model comparison
GPT-4o vs GPT-5.1
GPT-5.1 is the stronger model overall, scoring 48.8 to 28.5 on the Noometry Index.
Last verified . 52 shared benchmarks.
Summary
- They share 52 benchmarks with published results for both. GPT-4o scores higher in 0 categories and GPT-5.1 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.1 leads 50.3 to 10.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.4% for GPT-4o and 88.6% for GPT-5.1.
- GPT-5.1 is cheaper at $1.25 / $10 per million input/output tokens, against $2.50 / $10 for GPT-4o.
- GPT-5.1 accepts more context: 400K tokens versus 128K.
Side by side
| GPT-4o | GPT-5.1 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 28.5 | 48.8 |
| Released | 2024-05-13 | 2025-11-13 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 400K |
| Max output | 16K | 128K |
| Input $ / M tokens | $2.50 | $1.25 |
| Output $ / M tokens | $10 | $10 |
| Results tracked | 72 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.1 leads
GPT-4o: 25.1 (#323), GPT-5.1: 46.9 (#59)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| SWE-bench Verified | 31% | 68% |
| SWE-bench Verified (bash only) | 21.6% | 66% |
| GSO | 0% | 13.7% |
| WeirdML | 25.1% | 60.8% |
| LiveBench Coding | 51.4% | 72.5% |
| LMArena Coding | 1297 | 1454 |
| Aider Polyglot | 45.3% | — |
| LMArena WebDev | — | 1395 |
| SciCode | — | 43.3% |
| BigCodeBench Instruct | 51.1% | — |
| BigCodeBench Complete | 61.1% | — |
| CadEval | 26% | — |
| ALE-Bench | — | 1,192 |
| HumanEval+ | 87.2% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use GPT-5.1 leads
GPT-4o: 20.8 (#133), GPT-5.1: 32.5 (#55)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LMArena Search | 1006 | 1199 |
| Terminal-Bench | — | 47.6% |
| GDPval | 9.9% | — |
| TheAgentCompany | 8.6% | — |
| Cybench | 12.5% | — |
| DeepResearch Bench | — | 42.8% |
| BALROG | 32.3% | — |
| METR Time Horizons | 40.8% | — |
| Vending-Bench 2 | — | 1,473 |
Reasoning GPT-5.1 leads
GPT-4o: 9.4 (#338), GPT-5.1: 39.9 (#50)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| ARC-AGI-2 | 0% | 17.6% |
| SimpleBench | 17.8% | 53.2% |
| ARC-AGI-1 | 4.5% | 72.8% |
| CritPt | 0% | 4.9% |
| Chess Puzzles | 13% | 32% |
| EnigmaEval | 0.8% | 11.2% |
| LiveBench Reasoning | 55.8% | 95.8% |
| LMArena Hard Prompts | 1281 | 1457 |
| DTBench | 64.5% | 90.1% |
| LiveBench Data Analysis | 60.9% | 72.1% |
| LMCA | 16.6% | 43.9% |
| Epoch Capabilities Index | 128.97 | 149.66 |
| ForecastBench | 57.7 | 58.1 |
| LiveBench | 55.3% | 78.8% |
| Mystery Game Puzzles | — | 19% |
Math GPT-5.1 leads
GPT-4o: 10.2 (#303), GPT-5.1: 50.3 (#45)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.4% | 88.6% |
| Omni-MATH | 29.3% | 46.4% |
| LiveBench Math | 49.5% | 94.5% |
| LMArena Math | 1285 | 1446 |
| FrontierMath (Feb 2025 set) | 0.3% | 31% |
| FrontierMath (Tiers 1-3) | 0.4% | — |
| MATH Level 5 | 53.3% | — |
| FrontierMath Tier 4 (v1) | — | 12.5% |
Knowledge GPT-5.1 leads
GPT-4o: 28.8 (#235), GPT-5.1: 50.5 (#65)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 49.2% | 87.6% |
| Humanity's Last Exam | 2.7% | 23.7% |
| SimpleQA Verified | 26% | 48% |
| MMLU-Pro | 71.3% | 57.9% |
| Vectara Hallucination Rate | 9.6% | 10.9% |
| GPQA (HELM) | 52% | 44.2% |
| LMArena Expert | 1250 | 1470 |
| Confabulations | 15.3% | — |
| MMLU | 88.1% | — |
Multimodal GPT-5.1 leads
GPT-4o: 33.9 (#85), GPT-5.1: 44.0 (#17)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LMArena Vision | 1137 | 1250 |
| VPCT | 40% | 58.7% |
| Video-MME | 71.9% | — |
| GeoBench | 71% | — |
| LMArena Document | — | 1403 |
| ScienceQA | 88.5% | — |
Multilingual GPT-5.1 leads
GPT-4o: 43.4 (#181), GPT-5.1: 54.0 (#51)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1283 | 1431 |
| LMArena Chinese | 1277 | 1495 |
| LMArena French | 1303 | 1451 |
| LMArena German | 1282 | 1439 |
| LMArena Japanese | 1257 | 1453 |
| LMArena Korean | 1234 | 1401 |
| LMArena Russian | 1286 | 1435 |
| LMArena Spanish | 1292 | 1430 |
Instruction Following GPT-5.1 leads
GPT-4o: 66.8 (#202), GPT-5.1: 84.0 (#1)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LiveBench Instruction Following | 68.6% | 93.3% |
| IFEval | 81.7% | 93.5% |
| LMArena Instruction Following | 1278 | 1442 |
Long Context GPT-5.1 leads
GPT-4o: 39.4 (#174), GPT-5.1: 47.6 (#14)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LMArena Longer Query | 1289 | 1448 |
| Fiction.LiveBench | 66.7% | — |
| CL-bench | — | 23.7% |
| CL-bench Life | — | 17.3% |
Writing & Preference GPT-5.1 leads
GPT-4o: 52.8 (#160), GPT-5.1: 64.7 (#53)
| Benchmark | GPT-4o | GPT-5.1 |
|---|---|---|
| LMArena Text | 1300 | 1443 |
| LMArena Creative Writing | 1291 | 1427 |
| WildBench | 82.8% | 86.3% |
| LMArena Multi-Turn | 1302 | 1449 |
| LiveBench Language | 47.6% | 80.2% |
| Short-Story Creative Writing | 81.8% | — |
Frequently asked questions
Is GPT-4o better than GPT-5.1?
GPT-5.1 is the stronger model overall, scoring 48.8 to 28.5 on the Noometry Index.
Which is cheaper, GPT-4o or GPT-5.1?
GPT-5.1 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-4o lists at $2.50 and $10.
Is GPT-4o or GPT-5.1 better for coding?
GPT-5.1 scores higher on coding benchmarks: 46.9 versus 25.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.1 does, with 400K tokens against 128K.
How many benchmarks do GPT-4o and GPT-5.1 share?
52 benchmarks have published results for both models. GPT-4o has 72 scored results on Noometry and GPT-5.1 has 63.