Model comparison
GPT-4o mini vs o1
o1 is the stronger model overall, scoring 40.9 to 25.5 on the Noometry Index. GPT-4o mini costs 100× 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. GPT-4o mini scores higher in 1 category and o1 in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where o1 leads 36.1 to 10.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.9% for GPT-4o mini and 73.3% for o1.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $15 / $60 for o1.
- o1 accepts more context: 200K tokens versus 128K.
Side by side
| GPT-4o mini | o1 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 25.5 | 40.9 |
| Released | 2024-07-18 | 2024-09-12 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 200K |
| Max output | 16K | 100K |
| Input $ / M tokens | $0.15 | $15 |
| Output $ / M tokens | $0.60 | $60 |
| Results tracked | 60 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o1 leads
GPT-4o mini: 22.0 (#335), o1: 46.1 (#70)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| Aider Polyglot | 3.6% | 61.7% |
| WeirdML | 11.8% | 47.6% |
| LiveBench Coding | 43.1% | 69.7% |
| LMArena Coding | 1290 | 1367 |
| HumanEval+ | 83.5% | 89% |
| MBPP+ | 72.2% | 80.2% |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 57.4% | — |
| CadEval | — | 56% |
Agentic & Tool Use GPT-4o mini leads
GPT-4o mini: 27.5 (#101), o1: 24.6 (#117)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| Cybench | — | 10% |
| BALROG | 17.4% | — |
| METR Time Horizons | — | 51.1% |
Reasoning o1 leads
GPT-4o mini: 8.7 (#347), o1: 27.9 (#111)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| SimpleBench | 10.7% | 41.7% |
| Chess Puzzles | 0% | 15% |
| LiveBench Reasoning | 32.8% | 91.6% |
| LMArena Hard Prompts | 1267 | 1371 |
| DTBench | 54.4% | 74.7% |
| LiveBench Data Analysis | 50% | 65.5% |
| LMCA | 10.4% | 22.3% |
| Epoch Capabilities Index | 126.56 | 141.91 |
| LiveBench | 41.3% | 75.7% |
| ARC-AGI-2 | 0% | — |
| Kagi LLM Benchmark | 28.8% | — |
| ARC-AGI-1 | — | 30.7% |
| EnigmaEval | — | 5.7% |
| Mystery Game Puzzles | 12% | — |
| PIQA | 88.7% | — |
Math o1 leads
GPT-4o mini: 10.4 (#314), o1: 36.1 (#175)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 14.7% |
| OTIS Mock AIME 2024-2025 | 6.9% | 73.3% |
| LiveBench Math | 36.3% | 80.3% |
| LMArena Math | 1267 | 1388 |
| MATH Level 5 | 52.6% | 94.7% |
| Omni-MATH | 28% | — |
| FrontierMath (Feb 2025 set) | — | 9.3% |
| GSM8K | 91.3% | — |
Knowledge o1 leads
GPT-4o mini: 17.7 (#284), o1: 41.5 (#110)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| GPQA Diamond | 37.7% | 76.8% |
| SimpleQA Verified | 8.3% | 41.1% |
| Confabulations | 37.2% | 11.7% |
| LMArena Expert | 1235 | 1361 |
| Humanity's Last Exam | — | 8% |
| MMLU-Pro | 60.3% | — |
| GPQA (HELM) | 36.8% | — |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal o1 leads
GPT-4o mini: 25.9 (#122), o1: 34.2 (#93)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| LMArena Vision | 1066 | 1168 |
| GeoBench | 64% | 80% |
| VPCT | 34% | 37% |
| Video-MME | 64.8% | — |
| SpatialViz-Bench | — | 41.4% |
Multilingual o1 leads
GPT-4o mini: 42.0 (#199), o1: 48.6 (#142)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| LMArena Non-English | 1266 | 1358 |
| LMArena Chinese | 1265 | 1394 |
| LMArena French | 1297 | 1344 |
| LMArena German | 1272 | 1337 |
| LMArena Japanese | 1216 | 1346 |
| LMArena Korean | 1195 | 1396 |
| LMArena Russian | 1275 | 1356 |
| LMArena Spanish | 1276 | 1345 |
Instruction Following o1 leads
GPT-4o mini: 61.9 (#239), o1: 74.8 (#86)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| LiveBench Instruction Following | 56.8% | 81.5% |
| LMArena Instruction Following | 1258 | 1367 |
| IFEval | 78.2% | — |
Long Context o1 leads
GPT-4o mini: 39.1 (#186), o1: 50.3 (#9)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| LMArena Longer Query | 1289 | 1378 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference o1 leads
GPT-4o mini: 39.5 (#248), o1: 55.6 (#144)
| Benchmark | GPT-4o mini | o1 |
|---|---|---|
| LMArena Text | 1286 | 1366 |
| LMArena Creative Writing | 1268 | 1348 |
| Short-Story Creative Writing | 67.2% | 70.2% |
| LMArena Multi-Turn | 1285 | 1369 |
| LiveBench Language | 28.6% | 65.4% |
| EQ-Bench Creative Writing | 873 | — |
| WildBench | 79.1% | — |
Frequently asked questions
Is GPT-4o mini better than o1?
o1 is the stronger model overall, scoring 40.9 to 25.5 on the Noometry Index. GPT-4o mini costs 100× less per token, which makes it the better buy when o1's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini or o1?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; o1 lists at $15 and $60.
Is GPT-4o mini or o1 better for coding?
o1 scores higher on coding benchmarks: 46.1 versus 22.0 in the Noometry coding category.
Which has the bigger context window?
o1 does, with 200K tokens against 128K.
How many benchmarks do GPT-4o mini and o1 share?
43 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and o1 has 52.