Model comparison
GPT-5 Mini vs o3
o3 is the stronger model overall, scoring 47.5 to 41.8 on the Noometry Index. GPT-5 Mini costs 5.1× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 52 shared benchmarks.
Summary
- They share 52 benchmarks with published results for both. GPT-5 Mini scores higher in 1 category and o3 in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3 leads 53.3 to 41.9.
- The biggest single-benchmark swing is SimpleQA Verified: 21.6% for GPT-5 Mini and 49.4% for o3.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $2 / $8 for o3.
- GPT-5 Mini accepts more context: 400K tokens versus 200K.
Side by side
| GPT-5 Mini | o3 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 41.8 | 47.5 |
| Released | 2025-08-07 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $0.25 | $2 |
| Output $ / M tokens | $2 | $8 |
| Results tracked | 60 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
GPT-5 Mini: 40.1 (#146), o3: 46.8 (#64)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| SWE-bench Verified | 64.7% | 62.3% |
| SWE-bench Verified (bash only) | 59.8% | 58.4% |
| WeirdML | 52.7% | 52.4% |
| LMArena Coding | 1406 | 1408 |
| ALE-Bench | 799.77 | 933.55 |
| Aider Polyglot | — | 81.3% |
| SWE-bench Multilingual | 39.7% | — |
| SciCode | 39.2% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
| AlgoTune | 1.38 | — |
Agentic & Tool Use o3 leads
GPT-5 Mini: 31.1 (#70), o3: 34.5 (#44)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 55.5% | 63% |
| Terminal-Bench | 34.8% | — |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | -31.18 | — |
Reasoning o3 leads
GPT-5 Mini: 23.9 (#168), o3: 32.0 (#78)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| ARC-AGI-2 | 4.4% | 6.5% |
| Kagi LLM Benchmark | 70.3% | 67.6% |
| ARC-AGI-1 | 54.3% | 60.8% |
| CritPt | 0% | 1.4% |
| Chess Puzzles | 30% | 38% |
| EnigmaEval | 8.2% | 13.1% |
| LMArena Hard Prompts | 1380 | 1402 |
| Mystery Game Puzzles | 10% | 29% |
| DTBench | 80.5% | 84.8% |
| LMCA | 34.2% | 39.7% |
| Epoch Capabilities Index | 145.52 | 146.86 |
| ForecastBench | 61 | 62.5 |
| SimpleBench | — | 53.1% |
Math o3 leads
GPT-5 Mini: 46.7 (#69), o3: 50.2 (#58)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 46.7% | 33.3% |
| OTIS Mock AIME 2024-2025 | 86.7% | 84.4% |
| Omni-MATH | 72.2% | 71.4% |
| LMArena Math | 1378 | 1426 |
| MATH Level 5 | 97.8% | 97.8% |
| FrontierMath (Feb 2025 set) | 27.2% | 18.7% |
| FrontierMath Tier 4 (v1) | 6.3% | 2.1% |
| FrontierMath Tier 4 | 12.2% | — |
| ProofBench | 9% | — |
Knowledge o3 leads
GPT-5 Mini: 45.6 (#86), o3: 54.6 (#52)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| GPQA Diamond | 75% | 81.8% |
| Humanity's Last Exam | 19.4% | 20.3% |
| SimpleQA Verified | 21.6% | 49.4% |
| MMLU-Pro | 83.5% | 85.9% |
| Confabulations | 13.3% | 14.4% |
| GPQA (HELM) | 75.6% | 75.3% |
| LMArena Expert | 1379 | 1402 |
| Vectara Hallucination Rate | 12.9% | — |
Multimodal o3 leads
GPT-5 Mini: 35.6 (#85), o3: 41.4 (#36)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| LMArena Vision | 1202 | 1214 |
| VPCT | 40.2% | 52% |
| GeoBench | — | 74% |
Multilingual o3 leads
GPT-5 Mini: 48.9 (#137), o3: 51.7 (#105)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| LMArena Non-English | 1363 | 1401 |
| LMArena Chinese | 1385 | 1437 |
| LMArena French | 1386 | 1430 |
| LMArena German | 1366 | 1420 |
| LMArena Japanese | 1341 | 1403 |
| LMArena Korean | 1308 | 1370 |
| LMArena Russian | 1362 | 1406 |
| LMArena Spanish | 1355 | 1395 |
Instruction Following GPT-5 Mini leads
GPT-5 Mini: 76.2 (#46), o3: 72.8 (#127)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| IFEval | 92.7% | 86.9% |
| LMArena Instruction Following | 1357 | 1368 |
Long Context o3 leads
GPT-5 Mini: 41.9 (#132), o3: 53.3 (#6)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| Fiction.LiveBench | 69.4% | 88.9% |
| LMArena Longer Query | 1355 | 1372 |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
GPT-5 Mini: 55.2 (#148), o3: 63.5 (#64)
| Benchmark | GPT-5 Mini | o3 |
|---|---|---|
| LMArena Text | 1373 | 1410 |
| LMArena Creative Writing | 1325 | 1359 |
| Short-Story Creative Writing | 83.1% | 83.9% |
| EQ-Bench Creative Writing | 1313 | 1676 |
| WildBench | 85.5% | 86.1% |
| LMArena Multi-Turn | 1363 | 1405 |
Frequently asked questions
Is GPT-5 Mini better than o3?
o3 is the stronger model overall, scoring 47.5 to 41.8 on the Noometry Index. GPT-5 Mini costs 5.1× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Mini or o3?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; o3 lists at $2 and $8.
Is GPT-5 Mini or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 40.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 200K.
How many benchmarks do GPT-5 Mini and o3 share?
52 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and o3 has 63.