Model comparison
GPT-4o mini vs o4-mini
o4-mini is the stronger model overall, scoring 41.6 to 25.5 on the Noometry Index. GPT-4o mini costs 7.3× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Last verified . 42 shared benchmarks.
Summary
- They share 42 benchmarks with published results for both. GPT-4o mini scores higher in 0 categories and o4-mini in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where o4-mini leads 40.8 to 10.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.9% for GPT-4o mini and 81.7% for o4-mini.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- o4-mini accepts more context: 200K tokens versus 128K.
Side by side
| GPT-4o mini | o4-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 25.5 | 41.6 |
| Released | 2024-07-18 | 2025-04-16 |
| 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 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
GPT-4o mini: 22.0 (#335), o4-mini: 40.9 (#127)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| Aider Polyglot | 3.6% | 72% |
| WeirdML | 11.8% | 52.6% |
| LMArena Coding | 1290 | 1368 |
| SWE-bench Verified (bash only) | — | 45% |
| GSO | — | 3.6% |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 43.1% | — |
| BigCodeBench Complete | 57.4% | — |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use o4-mini leads
GPT-4o mini: 27.5 (#101), o4-mini: 32.6 (#61)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
| BALROG | 17.4% | — |
| METR Time Horizons | — | 63.9% |
Reasoning o4-mini leads
GPT-4o mini: 8.7 (#347), o4-mini: 24.6 (#162)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| ARC-AGI-2 | 0% | 6.1% |
| SimpleBench | 10.7% | 38.7% |
| Kagi LLM Benchmark | 28.8% | 67.6% |
| Chess Puzzles | 0% | 26% |
| LMArena Hard Prompts | 1267 | 1351 |
| Mystery Game Puzzles | 12% | 5% |
| DTBench | 54.4% | 77.6% |
| LMCA | 10.4% | 26.5% |
| Epoch Capabilities Index | 126.56 | 145.64 |
| ARC-AGI-1 | — | 58.7% |
| CritPt | — | 0.6% |
| EnigmaEval | — | 9.2% |
| LiveBench Reasoning | 32.8% | — |
| LiveBench Data Analysis | 50% | — |
| ForecastBench | — | 61.8 |
| LiveBench | 41.3% | — |
| PIQA | 88.7% | — |
Math o4-mini leads
GPT-4o mini: 10.4 (#314), o4-mini: 40.8 (#89)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 36.1% |
| OTIS Mock AIME 2024-2025 | 6.9% | 81.7% |
| Omni-MATH | 28% | 72% |
| LMArena Math | 1267 | 1389 |
| MATH Level 5 | 52.6% | 97.8% |
| FrontierMath Tier 4 | — | 4.9% |
| LiveBench Math | 36.3% | — |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
| GSM8K | 91.3% | — |
Knowledge o4-mini leads
GPT-4o mini: 17.7 (#284), o4-mini: 43.6 (#91)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| GPQA Diamond | 37.7% | 79.6% |
| SimpleQA Verified | 8.3% | 19.6% |
| MMLU-Pro | 60.3% | 82% |
| Confabulations | 37.2% | 15.8% |
| GPQA (HELM) | 36.8% | 73.5% |
| LMArena Expert | 1235 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| Vectara Hallucination Rate | — | 18.6% |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal o4-mini leads
GPT-4o mini: 25.9 (#122), o4-mini: 40.2 (#49)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| LMArena Vision | 1066 | 1194 |
| GeoBench | 64% | 64% |
| VPCT | 34% | 57.5% |
| Video-MME | 64.8% | — |
Multilingual o4-mini leads
GPT-4o mini: 42.0 (#199), o4-mini: 47.0 (#154)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| LMArena Non-English | 1266 | 1337 |
| LMArena Chinese | 1265 | 1354 |
| LMArena French | 1297 | 1364 |
| LMArena German | 1272 | 1336 |
| LMArena Japanese | 1216 | 1308 |
| LMArena Korean | 1195 | 1312 |
| LMArena Russian | 1275 | 1334 |
| LMArena Spanish | 1276 | 1347 |
Instruction Following o4-mini leads
GPT-4o mini: 61.9 (#239), o4-mini: 75.2 (#68)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| IFEval | 78.2% | 92.8% |
| LMArena Instruction Following | 1258 | 1321 |
| LiveBench Instruction Following | 56.8% | — |
Long Context o4-mini leads
GPT-4o mini: 39.1 (#186), o4-mini: 45.5 (#33)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| LMArena Longer Query | 1289 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference o4-mini leads
GPT-4o mini: 39.5 (#248), o4-mini: 54.0 (#152)
| Benchmark | GPT-4o mini | o4-mini |
|---|---|---|
| LMArena Text | 1286 | 1353 |
| LMArena Creative Writing | 1268 | 1294 |
| Short-Story Creative Writing | 67.2% | 75% |
| WildBench | 79.1% | 85.4% |
| LMArena Multi-Turn | 1285 | 1350 |
| EQ-Bench Creative Writing | 873 | — |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than o4-mini?
o4-mini is the stronger model overall, scoring 41.6 to 25.5 on the Noometry Index. GPT-4o mini costs 7.3× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini or o4-mini?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is GPT-4o mini or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 40.9 versus 22.0 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 128K.
How many benchmarks do GPT-4o mini and o4-mini share?
42 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and o4-mini has 60.