Model comparison
GPT-5.1 vs GPT-5.2
GPT-5.2 is the stronger model overall, scoring 54.1 to 49.0 on the Noometry Index.
Last verified . 48 shared benchmarks.
Summary
- They share 48 benchmarks with published results for both. GPT-5.1 scores higher in 3 categories and GPT-5.2 in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.2 leads 50.2 to 39.8.
- The biggest single-benchmark swing is ARC-AGI-2: 17.6% for GPT-5.1 and 52.9% for GPT-5.2.
- GPT-5.1 is cheaper at $1.25 / $10 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
Side by side
| GPT-5.1 | GPT-5.2 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 49.0 | 54.1 |
| Released | 2025-11-13 | 2025-12-11 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 400K |
| Max output | 128K | 128K |
| Input $ / M tokens | $1.25 | $1.75 |
| Output $ / M tokens | $10 | $14 |
| Results tracked | 63 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.2 leads
GPT-5.1: 46.4 (#66), GPT-5.2: 51.6 (#37)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| SWE-bench Verified | 68% | 73.8% |
| SWE-bench Verified (bash only) | 66% | 72.8% |
| LMArena WebDev | 1395 | 1416 |
| GSO | 13.7% | 27.4% |
| WeirdML | 60.8% | 72.2% |
| LMArena Coding | 1454 | 1447 |
| ALE-Bench | 1,192 | 1,294 |
| SWE-bench Multilingual | — | 66.7% |
| SciCode | 43.3% | — |
| LiveBench Coding | 72.5% | — |
| AlgoTune | — | 2.05 |
Agentic & Tool Use GPT-5.2 leads
GPT-5.1: 32.7 (#60), GPT-5.2: 40.2 (#24)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| Terminal-Bench | 47.6% | 64.9% |
| DeepResearch Bench | 42.8% | 41.1% |
| LMArena Search | 1199 | 1207 |
| Vending-Bench 2 | 1,473 | 3,591 |
| Berkeley Function Calling Leaderboard | — | 55.9% |
| GDPval | — | 49.7% |
| Remote Labor Index | — | 2.5% |
| τ²-bench Airline | — | 83% |
| τ²-bench Banking | — | 32.2% |
| τ²-bench Retail | — | 81.6% |
| τ²-bench Telecom | — | 89.7% |
| METR Time Horizons | — | 75.3% |
Reasoning GPT-5.2 leads
GPT-5.1: 39.8 (#58), GPT-5.2: 50.2 (#35)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| ARC-AGI-2 | 17.6% | 52.9% |
| SimpleBench | 53.2% | 45.8% |
| ARC-AGI-1 | 72.8% | 86.2% |
| Chess Puzzles | 32% | 49% |
| EnigmaEval | 11.2% | 10.4% |
| LMArena Hard Prompts | 1457 | 1445 |
| Mystery Game Puzzles | 19% | 23% |
| DTBench | 90.1% | 90.9% |
| LMCA | 43.9% | 43.9% |
| Epoch Capabilities Index | 149.64 | 153.45 |
| ForecastBench | 58.1 | 60.1 |
| Kagi LLM Benchmark | — | 73.3% |
| NYT Connections (extended) | — | 83.6% |
| CritPt | 4.9% | — |
| EBR-Bench | — | 23% |
| LiveBench Reasoning | 95.8% | — |
| LiveBench Data Analysis | 72.1% | — |
| LiveBench | 78.8% | — |
Math GPT-5.2 leads
GPT-5.1: 52.2 (#51), GPT-5.2: 60.0 (#38)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.6% | 96.1% |
| LMArena Math | 1447 | 1440 |
| FrontierMath (Feb 2025 set) | 31% | 40.7% |
| FrontierMath Tier 4 (v1) | 12.5% | 18.8% |
| FrontierMath (Tiers 1-3) | — | 67.4% |
| FrontierMath Tier 4 | — | 31.7% |
| MathArena Final-Answer Competitions | — | 72% |
| ProofBench | — | 15% |
| Omni-MATH | 46.4% | — |
| LiveBench Math | 94.5% | — |
Knowledge GPT-5.2 leads
GPT-5.1: 50.6 (#71), GPT-5.2: 59.3 (#32)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| GPQA Diamond | 87.6% | 91.4% |
| Humanity's Last Exam | 23.7% | 27.8% |
| SimpleQA Verified | 48% | 37.1% |
| Vectara Hallucination Rate | 10.9% | 8.4% |
| LMArena Expert | 1470 | 1445 |
| MMLU-Pro | 57.9% | — |
| GPQA (HELM) | 44.2% | — |
Multimodal GPT-5.2 leads
GPT-5.1: 44.8 (#19), GPT-5.2: 51.3 (#7)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| LMArena Vision | 1250 | 1268 |
| VPCT | 58.7% | 84% |
| LMArena Document | 1403 | 1405 |
| Furniture Assembly | — | 38.3% |
Multilingual Too close to call
GPT-5.1: 53.8 (#56), GPT-5.2: 53.4 (#67)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| LMArena Non-English | 1431 | 1425 |
| LMArena Chinese | 1495 | 1460 |
| LMArena French | 1450 | 1455 |
| LMArena German | 1438 | 1448 |
| LMArena Japanese | 1453 | 1420 |
| LMArena Korean | 1401 | 1392 |
| LMArena Russian | 1435 | 1440 |
| LMArena Spanish | 1433 | 1433 |
Instruction Following GPT-5.1 leads
GPT-5.1: 83.9 (#1), GPT-5.2: 74.7 (#89)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| LMArena Instruction Following | 1443 | 1417 |
| LiveBench Instruction Following | 93.3% | — |
| IFEval | 93.5% | — |
Long Context GPT-5.1 leads
GPT-5.1: 47.6 (#14), GPT-5.2: 44.0 (#78)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| CL-bench | 23.7% | 18.2% |
| LMArena Longer Query | 1447 | 1428 |
| CL-bench Life | 17.3% | — |
Writing & Preference GPT-5.2 leads
GPT-5.1: 64.5 (#55), GPT-5.2: 66.8 (#32)
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| LMArena Text | 1443 | 1439 |
| LMArena Creative Writing | 1427 | 1401 |
| LMArena Multi-Turn | 1450 | 1458 |
| EQ-Bench Creative Writing | — | 1703 |
| WildBench | 86.3% | — |
| LiveBench Language | 80.2% | — |
Frequently asked questions
Is GPT-5.1 better than GPT-5.2?
GPT-5.2 is the stronger model overall, scoring 54.1 to 49.0 on the Noometry Index.
Which is cheaper, GPT-5.1 or GPT-5.2?
GPT-5.1 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-5.2 lists at $1.75 and $14.
Is GPT-5.1 or GPT-5.2 better for coding?
GPT-5.2 scores higher on coding benchmarks: 51.6 versus 46.4 in the Noometry coding category.
Which has the bigger context window?
Both accept 400K tokens.
How many benchmarks do GPT-5.1 and GPT-5.2 share?
48 benchmarks have published results for both models. GPT-5.1 has 63 scored results on Noometry and GPT-5.2 has 67.