Model comparison
o1 vs o3-mini
o1 is the stronger model overall, scoring 40.9 to 36.7 on the Noometry Index. o3-mini costs 14× less per token, which makes it the better buy when o1's lead doesn't matter for your workload.
Last verified . 43 shared benchmarks.
Summary
- They share 43 benchmarks with published results for both. o1 scores higher in 7 categories and o3-mini in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in long context, where o1 leads 50.3 to 33.8.
- The biggest single-benchmark swing is Fiction.LiveBench: 83.3% for o1 and 50% for o3-mini.
- o3-mini is cheaper at $1.10 / $4.40 per million input/output tokens, against $15 / $60 for o1.
Side by side
| o1 | o3-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 40.9 | 36.7 |
| Released | 2024-09-12 | 2024-12-20 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 200K |
| Max output | 100K | 100K |
| Input $ / M tokens | $15 | $1.10 |
| Output $ / M tokens | $60 | $4.40 |
| Results tracked | 52 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o1 leads
o1: 46.1 (#70), o3-mini: 40.8 (#132)
| Benchmark | o1 | o3-mini |
|---|---|---|
| Aider Polyglot | 61.7% | 60.4% |
| WeirdML | 47.6% | 43.7% |
| LiveBench Coding | 69.7% | 82.7% |
| LMArena Coding | 1367 | 1378 |
| CadEval | 56% | 54% |
| SciCode | — | 39.8% |
| GSO | — | 1.3% |
| HumanEval+ | 89% | — |
| MBPP+ | 80.2% | — |
Agentic & Tool Use o3-mini leads
o1: 24.6 (#117), o3-mini: 29.6 (#84)
| Benchmark | o1 | o3-mini |
|---|---|---|
| Cybench | 10% | 22.5% |
| METR Time Horizons | 51.1% | — |
Reasoning o1 leads
o1: 27.9 (#111), o3-mini: 16.3 (#305)
| Benchmark | o1 | o3-mini |
|---|---|---|
| SimpleBench | 41.7% | 22.8% |
| ARC-AGI-1 | 30.7% | 34.5% |
| Chess Puzzles | 15% | 17% |
| LiveBench Reasoning | 91.6% | 89.6% |
| LMArena Hard Prompts | 1371 | 1366 |
| DTBench | 74.7% | 68.8% |
| LiveBench Data Analysis | 65.5% | 70.6% |
| LMCA | 22.3% | 19% |
| Epoch Capabilities Index | 141.91 | 140.34 |
| LiveBench | 75.7% | 75.9% |
| ARC-AGI-2 | — | 3% |
| CritPt | — | 0.3% |
| EnigmaEval | 5.7% | — |
| Mystery Game Puzzles | — | 7% |
| ForecastBench | — | 59.6 |
Math o1 leads
o1: 36.1 (#175), o3-mini: 28.1 (#244)
| Benchmark | o1 | o3-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 14.7% | 18.6% |
| OTIS Mock AIME 2024-2025 | 73.3% | 76.9% |
| LiveBench Math | 80.3% | 77.3% |
| LMArena Math | 1388 | 1396 |
| MATH Level 5 | 94.7% | 96.5% |
| FrontierMath (Feb 2025 set) | 9.3% | 12.4% |
| FrontierMath Tier 4 | — | 0% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge o1 leads
o1: 41.5 (#110), o3-mini: 38.3 (#146)
| Benchmark | o1 | o3-mini |
|---|---|---|
| GPQA Diamond | 76.8% | 77% |
| SimpleQA Verified | 41.1% | 15.3% |
| Confabulations | 11.7% | 17.9% |
| LMArena Expert | 1361 | 1364 |
| Humanity's Last Exam | 8% | — |
Multimodal Not comparable
o1: 34.2 (#93), o3-mini: —
| Benchmark | o1 | o3-mini |
|---|---|---|
| LMArena Vision | 1168 | — |
| GeoBench | 80% | — |
| VPCT | 37% | — |
| SpatialViz-Bench | 41.4% | — |
Multilingual o1 leads
o1: 48.6 (#142), o3-mini: 45.7 (#164)
| Benchmark | o1 | o3-mini |
|---|---|---|
| LMArena Non-English | 1358 | 1319 |
| LMArena Chinese | 1394 | 1379 |
| LMArena French | 1344 | 1334 |
| LMArena German | 1337 | 1303 |
| LMArena Japanese | 1346 | 1286 |
| LMArena Korean | 1396 | 1314 |
| LMArena Russian | 1356 | 1304 |
| LMArena Spanish | 1345 | 1321 |
Instruction Following Too close to call
o1: 74.8 (#86), o3-mini: 75.1 (#72)
| Benchmark | o1 | o3-mini |
|---|---|---|
| LiveBench Instruction Following | 81.5% | 84.4% |
| LMArena Instruction Following | 1367 | 1337 |
Long Context o1 leads
o1: 50.3 (#9), o3-mini: 33.8 (#256)
| Benchmark | o1 | o3-mini |
|---|---|---|
| Fiction.LiveBench | 83.3% | 50% |
| LMArena Longer Query | 1378 | 1343 |
Writing & Preference o1 leads
o1: 55.6 (#144), o3-mini: 50.3 (#182)
| Benchmark | o1 | o3-mini |
|---|---|---|
| LMArena Text | 1366 | 1337 |
| LMArena Creative Writing | 1348 | 1286 |
| Short-Story Creative Writing | 70.2% | 61.7% |
| LMArena Multi-Turn | 1369 | 1320 |
| LiveBench Language | 65.4% | 50.7% |
Frequently asked questions
Is o1 better than o3-mini?
o1 is the stronger model overall, scoring 40.9 to 36.7 on the Noometry Index. o3-mini costs 14× less per token, which makes it the better buy when o1's lead doesn't matter for your workload.
Which is cheaper, o1 or o3-mini?
o3-mini is cheaper. It lists at $1.10 per million input tokens and $4.40 per million output tokens; o1 lists at $15 and $60.
Is o1 or o3-mini better for coding?
o1 scores higher on coding benchmarks: 46.1 versus 40.8 in the Noometry coding category.
Which has the bigger context window?
Both accept 200K tokens.
How many benchmarks do o1 and o3-mini share?
43 benchmarks have published results for both models. o1 has 52 scored results on Noometry and o3-mini has 51.