Model comparison
GPT-4 vs o4-mini
o4-mini is the stronger model overall, scoring 41.6 to 29.1 on the Noometry Index.
Last verified . 28 shared benchmarks.
Summary
- They share 28 benchmarks with published results for both. GPT-4 scores higher in 0 categories and o4-mini in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where o4-mini leads 40.8 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 81.7% for o4-mini.
- o4-mini is cheaper at $1.10 / $4.40 per million input/output tokens, against $30 / $60 for GPT-4.
- o4-mini accepts more context: 200K tokens versus 8K.
Side by side
| GPT-4 | o4-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 29.1 | 41.6 |
| Released | 2023-03-14 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 8K | 200K |
| Max output | 8K | 100K |
| Input $ / M tokens | $30 | $1.10 |
| Output $ / M tokens | $60 | $4.40 |
| Results tracked | 38 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
GPT-4: 31.6 (#283), o4-mini: 40.9 (#127)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| WeirdML | 12.4% | 52.6% |
| LMArena Coding | 1254 | 1368 |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| GSO | — | 3.6% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, o4-mini: 32.6 (#61)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| METR Time Horizons | 36.1% | 63.9% |
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
Reasoning o4-mini leads
GPT-4: 17.8 (#289), o4-mini: 24.6 (#162)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| Chess Puzzles | 4% | 26% |
| LMArena Hard Prompts | 1241 | 1351 |
| Mystery Game Puzzles | 12% | 5% |
| DTBench | 62.7% | 77.6% |
| LMCA | 17.1% | 26.5% |
| Epoch Capabilities Index | 125.89 | 145.64 |
| ForecastBench | 57.8 | 61.8 |
| ARC-AGI-2 | — | 6.1% |
| SimpleBench | — | 38.7% |
| Kagi LLM Benchmark | — | 67.6% |
| ARC-AGI-1 | — | 58.7% |
| CritPt | — | 0.6% |
| EnigmaEval | — | 9.2% |
| BIG-Bench Hard | 75.1% | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math o4-mini leads
GPT-4: 10.8 (#309), o4-mini: 40.8 (#89)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 81.7% |
| LMArena Math | 1269 | 1389 |
| MATH Level 5 | 23% | 97.8% |
| FrontierMath (Tiers 1-3) | — | 36.1% |
| FrontierMath Tier 4 | — | 4.9% |
| Omni-MATH | — | 72% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
| GSM8K | 92% | — |
Knowledge o4-mini leads
GPT-4: 18.4 (#282), o4-mini: 43.6 (#91)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| GPQA Diamond | 35.7% | 79.6% |
| LMArena Expert | 1211 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| SimpleQA Verified | — | 19.6% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multimodal Not comparable
GPT-4: —, o4-mini: 40.2 (#49)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| LMArena Vision | — | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
Multilingual o4-mini leads
GPT-4: 40.6 (#215), o4-mini: 47.0 (#154)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| LMArena Non-English | 1246 | 1337 |
| LMArena Chinese | 1242 | 1354 |
| LMArena French | 1283 | 1364 |
| LMArena German | 1251 | 1336 |
| LMArena Japanese | 1209 | 1308 |
| LMArena Korean | 1184 | 1312 |
| LMArena Russian | 1251 | 1334 |
| LMArena Spanish | 1261 | 1347 |
Instruction Following o4-mini leads
GPT-4: 65.3 (#222), o4-mini: 75.2 (#68)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1241 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
GPT-4: 37.7 (#212), o4-mini: 45.5 (#33)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| LMArena Longer Query | 1244 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference o4-mini leads
GPT-4: 34.9 (#268), o4-mini: 54.0 (#152)
| Benchmark | GPT-4 | o4-mini |
|---|---|---|
| LMArena Text | 1263 | 1353 |
| LMArena Creative Writing | 1244 | 1294 |
| LMArena Multi-Turn | 1257 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 752 | — |
| WildBench | — | 85.4% |
Frequently asked questions
Is GPT-4 better than o4-mini?
o4-mini is the stronger model overall, scoring 41.6 to 29.1 on the Noometry Index.
Which is cheaper, GPT-4 or o4-mini?
o4-mini is cheaper. It lists at $1.10 per million input tokens and $4.40 per million output tokens; GPT-4 lists at $30 and $60.
Is GPT-4 or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 40.9 versus 31.6 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 8K.
How many benchmarks do GPT-4 and o4-mini share?
28 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and o4-mini has 60.