Model comparison
GPT-5 Mini vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.1× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
StepFunSummary
- They share 17 benchmarks with published results for both. GPT-5 Mini scores higher in 3 categories and Step 5 Preview in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 5 Preview leads 40.0 to 23.9.
- The biggest single-benchmark swing is ProofBench: 9% for GPT-5 Mini and 42% for Step 5 Preview.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 400K.
Side by side
| GPT-5 Mini | Step 5 Preview | |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 41.8 | 47.9 |
| Released | 2025-08-07 | 2026-09-16 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.02M |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.25 | $1 |
| Output $ / M tokens | $2 | $2.70 |
| Results tracked | 60 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
GPT-5 Mini: 40.1 (#146), Step 5 Preview: 51.8 (#36)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| SciCode | 39.2% | 58.9% |
| LMArena Coding | 1406 | 1480 |
| SWE-bench Verified | 64.7% | — |
| SWE-bench Verified (bash only) | 59.8% | — |
| LMArena WebDev | — | 1564 |
| SWE-bench Multilingual | 39.7% | — |
| WeirdML | 52.7% | — |
| ALE-Bench | 799.77 | — |
| AlgoTune | 1.38 | — |
Agentic & Tool Use Not comparable
GPT-5 Mini: 31.1 (#70), Step 5 Preview: —
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| Terminal-Bench | 34.8% | — |
| Berkeley Function Calling Leaderboard | 55.5% | — |
| Vending-Bench 2 | -31.18 | — |
Reasoning Step 5 Preview leads
GPT-5 Mini: 23.9 (#168), Step 5 Preview: 40.0 (#57)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| CritPt | 0% | 20.9% |
| LMArena Hard Prompts | 1380 | 1465 |
| ARC-AGI-2 | 4.4% | — |
| Kagi LLM Benchmark | 70.3% | — |
| ARC-AGI-1 | 54.3% | — |
| Chess Puzzles | 30% | — |
| EnigmaEval | 8.2% | — |
| Mystery Game Puzzles | 10% | — |
| DTBench | 80.5% | — |
| LMCA | 34.2% | — |
| Epoch Capabilities Index | 145.52 | — |
| ForecastBench | 61 | — |
Math Too close to call
GPT-5 Mini: 46.7 (#69), Step 5 Preview: 46.1 (#72)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| ProofBench | 9% | 42% |
| LMArena Math | 1378 | 1470 |
| FrontierMath (Tiers 1-3) | 46.7% | — |
| FrontierMath Tier 4 | 12.2% | — |
| OTIS Mock AIME 2024-2025 | 86.7% | — |
| Omni-MATH | 72.2% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 27.2% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge GPT-5 Mini leads
GPT-5 Mini: 45.6 (#86), Step 5 Preview: 41.2 (#112)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1379 | 1470 |
| GPQA Diamond | 75% | — |
| Humanity's Last Exam | 19.4% | — |
| SimpleQA Verified | 21.6% | — |
| MMLU-Pro | 83.5% | — |
| Confabulations | 13.3% | — |
| Vectara Hallucination Rate | 12.9% | — |
| GPQA (HELM) | 75.6% | — |
Multimodal Step 5 Preview leads
GPT-5 Mini: 35.6 (#85), Step 5 Preview: 41.0 (#41)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Vision | 1202 | 1267 |
| VPCT | 40.2% | — |
Multilingual Step 5 Preview leads
GPT-5 Mini: 48.9 (#137), Step 5 Preview: 53.8 (#54)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1363 | 1432 |
| LMArena Chinese | 1385 | 1519 |
| LMArena Russian | 1362 | 1436 |
| LMArena Spanish | 1355 | 1439 |
| LMArena French | 1386 | — |
| LMArena German | 1366 | — |
| LMArena Japanese | 1341 | — |
| LMArena Korean | 1308 | — |
Instruction Following Too close to call
GPT-5 Mini: 76.2 (#46), Step 5 Preview: 76.0 (#50)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1357 | 1444 |
| IFEval | 92.7% | — |
Long Context Step 5 Preview leads
GPT-5 Mini: 41.9 (#132), Step 5 Preview: 44.5 (#64)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1355 | 1455 |
| Fiction.LiveBench | 69.4% | — |
Writing & Preference Step 5 Preview leads
GPT-5 Mini: 55.2 (#148), Step 5 Preview: 62.8 (#71)
| Benchmark | GPT-5 Mini | Step 5 Preview |
|---|---|---|
| LMArena Text | 1373 | 1442 |
| LMArena Creative Writing | 1325 | 1410 |
| LMArena Multi-Turn | 1363 | 1447 |
| Short-Story Creative Writing | 83.1% | — |
| EQ-Bench Creative Writing | 1313 | — |
| WildBench | 85.5% | — |
Frequently asked questions
Is GPT-5 Mini better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.1× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Mini or Step 5 Preview?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is GPT-5 Mini or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 40.1 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 400K.
How many benchmarks do GPT-5 Mini and Step 5 Preview share?
17 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and Step 5 Preview has 18.