Model comparison
GPT-5.2 vs o1
GPT-5.2 is the stronger model overall, scoring 54.1 to 40.9 on the Noometry Index.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. GPT-5.2 scores higher in 8 categories and o1 in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.2 leads 60.0 to 36.1.
- The biggest single-benchmark swing is ARC-AGI-1: 86.2% for GPT-5.2 and 30.7% for o1.
- GPT-5.2 is cheaper at $1.75 / $14 per million input/output tokens, against $15 / $60 for o1.
- GPT-5.2 accepts more context: 400K tokens versus 200K.
Side by side
| GPT-5.2 | o1 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 54.1 | 40.9 |
| Released | 2025-12-11 | 2024-09-12 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $1.75 | $15 |
| Output $ / M tokens | $14 | $60 |
| Results tracked | 67 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.2 leads
GPT-5.2: 51.6 (#37), o1: 46.1 (#70)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| WeirdML | 72.2% | 47.6% |
| LMArena Coding | 1447 | 1367 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 72.8% | — |
| Aider Polyglot | — | 61.7% |
| LMArena WebDev | 1416 | — |
| SWE-bench Multilingual | 66.7% | — |
| GSO | 27.4% | — |
| LiveBench Coding | — | 69.7% |
| CadEval | — | 56% |
| ALE-Bench | 1,294 | — |
| AlgoTune | 2.05 | — |
| HumanEval+ | — | 89% |
| MBPP+ | — | 80.2% |
Agentic & Tool Use GPT-5.2 leads
GPT-5.2: 40.2 (#24), o1: 24.6 (#117)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| METR Time Horizons | 75.3% | 51.1% |
| Terminal-Bench | 64.9% | — |
| 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% | — |
| Cybench | — | 10% |
| DeepResearch Bench | 41.1% | — |
| LMArena Search | 1207 | — |
| Vending-Bench 2 | 3,591 | — |
Reasoning GPT-5.2 leads
GPT-5.2: 50.2 (#35), o1: 27.9 (#111)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| SimpleBench | 45.8% | 41.7% |
| ARC-AGI-1 | 86.2% | 30.7% |
| Chess Puzzles | 49% | 15% |
| EnigmaEval | 10.4% | 5.7% |
| LMArena Hard Prompts | 1445 | 1371 |
| DTBench | 90.9% | 74.7% |
| LMCA | 43.9% | 22.3% |
| Epoch Capabilities Index | 153.45 | 141.91 |
| ARC-AGI-2 | 52.9% | — |
| Kagi LLM Benchmark | 73.3% | — |
| NYT Connections (extended) | 83.6% | — |
| EBR-Bench | 23% | — |
| LiveBench Reasoning | — | 91.6% |
| Mystery Game Puzzles | 23% | — |
| LiveBench Data Analysis | — | 65.5% |
| ForecastBench | 60.1 | — |
| LiveBench | — | 75.7% |
Math GPT-5.2 leads
GPT-5.2: 60.0 (#38), o1: 36.1 (#175)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 14.7% |
| OTIS Mock AIME 2024-2025 | 96.1% | 73.3% |
| LMArena Math | 1440 | 1388 |
| FrontierMath (Feb 2025 set) | 40.7% | 9.3% |
| FrontierMath Tier 4 | 31.7% | — |
| MathArena Final-Answer Competitions | 72% | — |
| ProofBench | 15% | — |
| LiveBench Math | — | 80.3% |
| MATH Level 5 | — | 94.7% |
| FrontierMath Tier 4 (v1) | 18.8% | — |
Knowledge GPT-5.2 leads
GPT-5.2: 59.3 (#32), o1: 41.5 (#110)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| GPQA Diamond | 91.4% | 76.8% |
| Humanity's Last Exam | 27.8% | 8% |
| SimpleQA Verified | 37.1% | 41.1% |
| LMArena Expert | 1445 | 1361 |
| Confabulations | — | 11.7% |
| Vectara Hallucination Rate | 8.4% | — |
Multimodal GPT-5.2 leads
GPT-5.2: 51.3 (#7), o1: 34.2 (#93)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| LMArena Vision | 1268 | 1168 |
| VPCT | 84% | 37% |
| GeoBench | — | 80% |
| Furniture Assembly | 38.3% | — |
| LMArena Document | 1405 | — |
| SpatialViz-Bench | — | 41.4% |
Multilingual GPT-5.2 leads
GPT-5.2: 53.4 (#67), o1: 48.6 (#142)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| LMArena Non-English | 1425 | 1358 |
| LMArena Chinese | 1460 | 1394 |
| LMArena French | 1455 | 1344 |
| LMArena German | 1448 | 1337 |
| LMArena Japanese | 1420 | 1346 |
| LMArena Korean | 1392 | 1396 |
| LMArena Russian | 1440 | 1356 |
| LMArena Spanish | 1433 | 1345 |
Instruction Following Too close to call
GPT-5.2: 74.7 (#89), o1: 74.8 (#86)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| LMArena Instruction Following | 1417 | 1367 |
| LiveBench Instruction Following | — | 81.5% |
Long Context o1 leads
GPT-5.2: 44.0 (#78), o1: 50.3 (#9)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| LMArena Longer Query | 1428 | 1378 |
| Fiction.LiveBench | — | 83.3% |
| CL-bench | 18.2% | — |
Writing & Preference GPT-5.2 leads
GPT-5.2: 66.8 (#32), o1: 55.6 (#144)
| Benchmark | GPT-5.2 | o1 |
|---|---|---|
| LMArena Text | 1439 | 1366 |
| LMArena Creative Writing | 1401 | 1348 |
| LMArena Multi-Turn | 1458 | 1369 |
| Short-Story Creative Writing | — | 70.2% |
| EQ-Bench Creative Writing | 1703 | — |
| LiveBench Language | — | 65.4% |
Frequently asked questions
Is GPT-5.2 better than o1?
GPT-5.2 is the stronger model overall, scoring 54.1 to 40.9 on the Noometry Index.
Which is cheaper, GPT-5.2 or o1?
GPT-5.2 is cheaper. It lists at $1.75 per million input tokens and $14 per million output tokens; o1 lists at $15 and $60.
Is GPT-5.2 or o1 better for coding?
GPT-5.2 scores higher on coding benchmarks: 51.6 versus 46.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.2 does, with 400K tokens against 200K.
How many benchmarks do GPT-5.2 and o1 share?
34 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and o1 has 52.