Model comparison
GPT-4o mini vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 25.5 on the Noometry Index.
Last verified . 17 shared benchmarks.
StepFunSummary
- They share 17 benchmarks with published results for both. GPT-4o mini scores higher in 0 categories and Step 3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Step 3 leads 37.6 to 10.4.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 28.8% for GPT-4o mini and 62.3% for Step 3.
- Step 3 has downloadable open weights; the other is API-only.
Side by side
| GPT-4o mini | Step 3 | |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 25.5 | 40.5 |
| Released | 2024-07-18 | — |
| Weights | Proprietary | Open |
| Context window | 128K | — |
| Max output | 16K | — |
| Input $ / M tokens | $0.15 | — |
| Output $ / M tokens | $0.60 | — |
| Results tracked | 60 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
GPT-4o mini: 22.0 (#335), Step 3: 40.1 (#147)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Coding | 1290 | 1367 |
| Aider Polyglot | 3.6% | — |
| WeirdML | 11.8% | — |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 43.1% | — |
| BigCodeBench Complete | 57.4% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use Not comparable
GPT-4o mini: 27.5 (#101), Step 3: —
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| BALROG | 17.4% | — |
Reasoning Step 3 leads
GPT-4o mini: 8.7 (#347), Step 3: 28.4 (#105)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 28.8% | 62.3% |
| LMArena Hard Prompts | 1267 | 1355 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 10.7% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 32.8% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 54.4% | — |
| LiveBench Data Analysis | 50% | — |
| LMCA | 10.4% | — |
| Epoch Capabilities Index | 126.56 | — |
| LiveBench | 41.3% | — |
| PIQA | 88.7% | — |
Math Step 3 leads
GPT-4o mini: 10.4 (#314), Step 3: 37.6 (#148)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Math | 1267 | 1366 |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| OTIS Mock AIME 2024-2025 | 6.9% | — |
| Omni-MATH | 28% | — |
| LiveBench Math | 36.3% | — |
| MATH Level 5 | 52.6% | — |
| GSM8K | 91.3% | — |
Knowledge Step 3 leads
GPT-4o mini: 17.7 (#284), Step 3: 36.8 (#164)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Expert | 1235 | 1333 |
| GPQA Diamond | 37.7% | — |
| SimpleQA Verified | 8.3% | — |
| MMLU-Pro | 60.3% | — |
| Confabulations | 37.2% | — |
| GPQA (HELM) | 36.8% | — |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal Step 3 leads
GPT-4o mini: 25.9 (#122), Step 3: 35.5 (#86)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Vision | 1066 | 1177 |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
Multilingual Step 3 leads
GPT-4o mini: 42.0 (#199), Step 3: 46.3 (#159)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Non-English | 1266 | 1327 |
| LMArena Chinese | 1265 | 1397 |
| LMArena German | 1272 | 1371 |
| LMArena Korean | 1195 | 1269 |
| LMArena Russian | 1275 | 1331 |
| LMArena Spanish | 1276 | 1371 |
| LMArena French | 1297 | — |
| LMArena Japanese | 1216 | — |
Instruction Following Step 3 leads
GPT-4o mini: 61.9 (#239), Step 3: 70.4 (#164)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1258 | 1332 |
| LiveBench Instruction Following | 56.8% | — |
| IFEval | 78.2% | — |
Long Context Step 3 leads
GPT-4o mini: 39.1 (#186), Step 3: 40.3 (#157)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Longer Query | 1289 | 1326 |
Writing & Preference Step 3 leads
GPT-4o mini: 39.5 (#248), Step 3: 54.3 (#151)
| Benchmark | GPT-4o mini | Step 3 |
|---|---|---|
| LMArena Text | 1286 | 1350 |
| LMArena Creative Writing | 1268 | 1321 |
| LMArena Multi-Turn | 1285 | 1341 |
| Short-Story Creative Writing | 67.2% | — |
| EQ-Bench Creative Writing | 873 | — |
| WildBench | 79.1% | — |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 25.5 on the Noometry Index.
Is GPT-4o mini or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 22.0 in the Noometry coding category.
How many benchmarks do GPT-4o mini and Step 3 share?
17 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Step 3 has 17.