Model comparison
GPT-4o mini vs o3-mini
o3-mini is the stronger model overall, scoring 36.7 to 25.5 on the Noometry Index. GPT-4o mini costs 7.3× less per token, which makes it the better buy when o3-mini's lead doesn't matter for your workload.
Last verified . 40 shared benchmarks.
Summary
- They share 40 benchmarks with published results for both. GPT-4o mini scores higher in 1 category and o3-mini in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where o3-mini leads 38.3 to 17.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.9% for GPT-4o mini and 76.9% for o3-mini.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $1.10 / $4.40 for o3-mini.
- o3-mini accepts more context: 200K tokens versus 128K.
Side by side
| GPT-4o mini | o3-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 25.5 | 36.7 |
| Released | 2024-07-18 | 2024-12-20 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 200K |
| Max output | 16K | 100K |
| Input $ / M tokens | $0.15 | $1.10 |
| Output $ / M tokens | $0.60 | $4.40 |
| Results tracked | 60 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3-mini leads
GPT-4o mini: 22.0 (#335), o3-mini: 40.8 (#132)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| Aider Polyglot | 3.6% | 60.4% |
| WeirdML | 11.8% | 43.7% |
| LiveBench Coding | 43.1% | 82.7% |
| LMArena Coding | 1290 | 1378 |
| SciCode | — | 39.8% |
| GSO | — | 1.3% |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 57.4% | — |
| CadEval | — | 54% |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use o3-mini leads
GPT-4o mini: 27.5 (#101), o3-mini: 29.6 (#84)
Reasoning o3-mini leads
GPT-4o mini: 8.7 (#347), o3-mini: 16.3 (#305)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| ARC-AGI-2 | 0% | 3% |
| SimpleBench | 10.7% | 22.8% |
| Chess Puzzles | 0% | 17% |
| LiveBench Reasoning | 32.8% | 89.6% |
| LMArena Hard Prompts | 1267 | 1366 |
| Mystery Game Puzzles | 12% | 7% |
| DTBench | 54.4% | 68.8% |
| LiveBench Data Analysis | 50% | 70.6% |
| LMCA | 10.4% | 19% |
| Epoch Capabilities Index | 126.56 | 140.34 |
| LiveBench | 41.3% | 75.9% |
| Kagi LLM Benchmark | 28.8% | — |
| ARC-AGI-1 | — | 34.5% |
| CritPt | — | 0.3% |
| ForecastBench | — | 59.6 |
| PIQA | 88.7% | — |
Math o3-mini leads
GPT-4o mini: 10.4 (#314), o3-mini: 28.1 (#244)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 18.6% |
| OTIS Mock AIME 2024-2025 | 6.9% | 76.9% |
| LiveBench Math | 36.3% | 77.3% |
| LMArena Math | 1267 | 1396 |
| MATH Level 5 | 52.6% | 96.5% |
| FrontierMath Tier 4 | — | 0% |
| Omni-MATH | 28% | — |
| FrontierMath (Feb 2025 set) | — | 12.4% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
| GSM8K | 91.3% | — |
Knowledge o3-mini leads
GPT-4o mini: 17.7 (#284), o3-mini: 38.3 (#146)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| GPQA Diamond | 37.7% | 77% |
| SimpleQA Verified | 8.3% | 15.3% |
| Confabulations | 37.2% | 17.9% |
| LMArena Expert | 1235 | 1364 |
| MMLU-Pro | 60.3% | — |
| GPQA (HELM) | 36.8% | — |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal Not comparable
GPT-4o mini: 25.9 (#122), o3-mini: —
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
Multilingual o3-mini leads
GPT-4o mini: 42.0 (#199), o3-mini: 45.7 (#164)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| LMArena Non-English | 1266 | 1319 |
| LMArena Chinese | 1265 | 1379 |
| LMArena French | 1297 | 1334 |
| LMArena German | 1272 | 1303 |
| LMArena Japanese | 1216 | 1286 |
| LMArena Korean | 1195 | 1314 |
| LMArena Russian | 1275 | 1304 |
| LMArena Spanish | 1276 | 1321 |
Instruction Following o3-mini leads
GPT-4o mini: 61.9 (#239), o3-mini: 75.1 (#72)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| LiveBench Instruction Following | 56.8% | 84.4% |
| LMArena Instruction Following | 1258 | 1337 |
| IFEval | 78.2% | — |
Long Context GPT-4o mini leads
GPT-4o mini: 39.1 (#186), o3-mini: 33.8 (#256)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| LMArena Longer Query | 1289 | 1343 |
| Fiction.LiveBench | — | 50% |
Writing & Preference o3-mini leads
GPT-4o mini: 39.5 (#248), o3-mini: 50.3 (#182)
| Benchmark | GPT-4o mini | o3-mini |
|---|---|---|
| LMArena Text | 1286 | 1337 |
| LMArena Creative Writing | 1268 | 1286 |
| Short-Story Creative Writing | 67.2% | 61.7% |
| LMArena Multi-Turn | 1285 | 1320 |
| LiveBench Language | 28.6% | 50.7% |
| EQ-Bench Creative Writing | 873 | — |
| WildBench | 79.1% | — |
Frequently asked questions
Is GPT-4o mini better than o3-mini?
o3-mini is the stronger model overall, scoring 36.7 to 25.5 on the Noometry Index. GPT-4o mini costs 7.3× less per token, which makes it the better buy when o3-mini's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini or o3-mini?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; o3-mini lists at $1.10 and $4.40.
Is GPT-4o mini or o3-mini better for coding?
o3-mini scores higher on coding benchmarks: 40.8 versus 22.0 in the Noometry coding category.
Which has the bigger context window?
o3-mini does, with 200K tokens against 128K.
How many benchmarks do GPT-4o mini and o3-mini share?
40 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and o3-mini has 51.