Model comparison
GPT-4o mini vs GPT-5 Mini
GPT-5 Mini is the stronger model overall, scoring 41.8 to 25.5 on the Noometry Index. GPT-4o mini costs 2.6× less per token, which makes it the better buy when GPT-5 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 0 categories and GPT-5 Mini in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 Mini leads 46.7 to 10.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.9% for GPT-4o mini and 86.7% for GPT-5 Mini.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.25 / $2 for GPT-5 Mini.
- GPT-5 Mini accepts more context: 400K tokens versus 128K.
Side by side
| GPT-4o mini | GPT-5 Mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 25.5 | 41.8 |
| Released | 2024-07-18 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 400K |
| Max output | 16K | 128K |
| Input $ / M tokens | $0.15 | $0.25 |
| Output $ / M tokens | $0.60 | $2 |
| Results tracked | 60 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Mini leads
GPT-4o mini: 22.0 (#335), GPT-5 Mini: 40.1 (#146)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| WeirdML | 11.8% | 52.7% |
| LMArena Coding | 1290 | 1406 |
| SWE-bench Verified | — | 64.7% |
| SWE-bench Verified (bash only) | — | 59.8% |
| Aider Polyglot | 3.6% | — |
| SWE-bench Multilingual | — | 39.7% |
| SciCode | — | 39.2% |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 43.1% | — |
| BigCodeBench Complete | 57.4% | — |
| ALE-Bench | — | 799.77 |
| AlgoTune | — | 1.38 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use GPT-5 Mini leads
GPT-4o mini: 27.5 (#101), GPT-5 Mini: 31.1 (#70)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | — | 34.8% |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| BALROG | 17.4% | — |
| Vending-Bench 2 | — | -31.18 |
Reasoning GPT-5 Mini leads
GPT-4o mini: 8.7 (#347), GPT-5 Mini: 23.9 (#168)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 0% | 4.4% |
| Kagi LLM Benchmark | 28.8% | 70.3% |
| Chess Puzzles | 0% | 30% |
| LMArena Hard Prompts | 1267 | 1380 |
| Mystery Game Puzzles | 12% | 10% |
| DTBench | 54.4% | 80.5% |
| LMCA | 10.4% | 34.2% |
| Epoch Capabilities Index | 126.56 | 145.52 |
| SimpleBench | 10.7% | — |
| ARC-AGI-1 | — | 54.3% |
| CritPt | — | 0% |
| EnigmaEval | — | 8.2% |
| LiveBench Reasoning | 32.8% | — |
| LiveBench Data Analysis | 50% | — |
| ForecastBench | — | 61 |
| LiveBench | 41.3% | — |
| PIQA | 88.7% | — |
Math GPT-5 Mini leads
GPT-4o mini: 10.4 (#314), GPT-5 Mini: 46.7 (#69)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 46.7% |
| OTIS Mock AIME 2024-2025 | 6.9% | 86.7% |
| Omni-MATH | 28% | 72.2% |
| LMArena Math | 1267 | 1378 |
| MATH Level 5 | 52.6% | 97.8% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 9% |
| LiveBench Math | 36.3% | — |
| FrontierMath (Feb 2025 set) | — | 27.2% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
| GSM8K | 91.3% | — |
Knowledge GPT-5 Mini leads
GPT-4o mini: 17.7 (#284), GPT-5 Mini: 45.6 (#86)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 37.7% | 75% |
| SimpleQA Verified | 8.3% | 21.6% |
| MMLU-Pro | 60.3% | 83.5% |
| Confabulations | 37.2% | 13.3% |
| GPQA (HELM) | 36.8% | 75.6% |
| LMArena Expert | 1235 | 1379 |
| Humanity's Last Exam | — | 19.4% |
| Vectara Hallucination Rate | — | 12.9% |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal GPT-5 Mini leads
GPT-4o mini: 25.9 (#122), GPT-5 Mini: 35.6 (#85)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1066 | 1202 |
| VPCT | 34% | 40.2% |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
Multilingual GPT-5 Mini leads
GPT-4o mini: 42.0 (#199), GPT-5 Mini: 48.9 (#137)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1266 | 1363 |
| LMArena Chinese | 1265 | 1385 |
| LMArena French | 1297 | 1386 |
| LMArena German | 1272 | 1366 |
| LMArena Japanese | 1216 | 1341 |
| LMArena Korean | 1195 | 1308 |
| LMArena Russian | 1275 | 1362 |
| LMArena Spanish | 1276 | 1355 |
Instruction Following GPT-5 Mini leads
GPT-4o mini: 61.9 (#239), GPT-5 Mini: 76.2 (#46)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| IFEval | 78.2% | 92.7% |
| LMArena Instruction Following | 1258 | 1357 |
| LiveBench Instruction Following | 56.8% | — |
Long Context GPT-5 Mini leads
GPT-4o mini: 39.1 (#186), GPT-5 Mini: 41.9 (#132)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1289 | 1355 |
| Fiction.LiveBench | — | 69.4% |
Writing & Preference GPT-5 Mini leads
GPT-4o mini: 39.5 (#248), GPT-5 Mini: 55.2 (#148)
| Benchmark | GPT-4o mini | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1286 | 1373 |
| LMArena Creative Writing | 1268 | 1325 |
| Short-Story Creative Writing | 67.2% | 83.1% |
| EQ-Bench Creative Writing | 873 | 1313 |
| WildBench | 79.1% | 85.5% |
| LMArena Multi-Turn | 1285 | 1363 |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than GPT-5 Mini?
GPT-5 Mini is the stronger model overall, scoring 41.8 to 25.5 on the Noometry Index. GPT-4o mini costs 2.6× less per token, which makes it the better buy when GPT-5 Mini's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini or GPT-5 Mini?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GPT-5 Mini lists at $0.25 and $2.
Is GPT-4o mini or GPT-5 Mini better for coding?
GPT-5 Mini scores higher on coding benchmarks: 40.1 versus 22.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 128K.
How many benchmarks do GPT-4o mini and GPT-5 Mini share?
40 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and GPT-5 Mini has 60.