Model comparison
GPT-5.4 nano vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 41.9 on the Noometry Index. GPT-5.4 nano costs 3.1× 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-5.4 nano scores higher in 1 category and Step 5 Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 5 Preview leads 40.0 to 23.7.
- The biggest single-benchmark swing is ProofBench: 5% for GPT-5.4 nano and 42% for Step 5 Preview.
- GPT-5.4 nano is cheaper at $0.20 / $1.25 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-5.4 nano | Step 5 Preview | |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 41.9 | 47.9 |
| Released | 2026-03-17 | 2026-09-16 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.02M |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.20 | $1 |
| Output $ / M tokens | $1.25 | $2.70 |
| Results tracked | 40 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
GPT-5.4 nano: 43.6 (#84), Step 5 Preview: 51.8 (#36)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| SciCode | 46.9% | 58.9% |
| LMArena Coding | 1405 | 1480 |
| LMArena WebDev | — | 1564 |
| WeirdML | 49.2% | — |
| ALE-Bench | 1,005 | — |
Reasoning Step 5 Preview leads
GPT-5.4 nano: 23.7 (#173), Step 5 Preview: 40.0 (#57)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| CritPt | 9.3% | 20.9% |
| LMArena Hard Prompts | 1381 | 1465 |
| ARC-AGI-2 | 5.7% | — |
| Kagi LLM Benchmark | 39.7% | — |
| ARC-AGI-1 | 51.5% | — |
| Chess Puzzles | 30% | — |
| Mystery Game Puzzles | 9% | — |
| DTBench | 80.3% | — |
| LMCA | 36.9% | — |
| Epoch Capabilities Index | 145.81 | — |
| ForecastBench | 57.3 | — |
Math Step 5 Preview leads
GPT-5.4 nano: 40.9 (#88), Step 5 Preview: 46.1 (#72)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| ProofBench | 5% | 42% |
| LMArena Math | 1406 | 1470 |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| FrontierMath (Feb 2025 set) | 25.9% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge Too close to call
GPT-5.4 nano: 41.9 (#103), Step 5 Preview: 41.2 (#112)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1396 | 1470 |
| GPQA Diamond | 78.5% | — |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 3.1% | — |
Multimodal Step 5 Preview leads
GPT-5.4 nano: 36.7 (#78), Step 5 Preview: 41.0 (#41)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Vision | 1196 | 1267 |
Multilingual Step 5 Preview leads
GPT-5.4 nano: 48.6 (#140), Step 5 Preview: 53.8 (#54)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1359 | 1432 |
| LMArena Chinese | 1392 | 1519 |
| LMArena Russian | 1363 | 1436 |
| LMArena Spanish | 1371 | 1439 |
| LMArena French | 1396 | — |
| LMArena German | 1367 | — |
| LMArena Japanese | 1343 | — |
| LMArena Korean | 1320 | — |
Instruction Following Step 5 Preview leads
GPT-5.4 nano: 71.9 (#144), Step 5 Preview: 76.0 (#50)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1362 | 1444 |
Long Context Step 5 Preview leads
GPT-5.4 nano: 41.6 (#137), Step 5 Preview: 44.5 (#64)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1366 | 1455 |
Writing & Preference Step 5 Preview leads
GPT-5.4 nano: 55.7 (#142), Step 5 Preview: 62.8 (#71)
| Benchmark | GPT-5.4 nano | Step 5 Preview |
|---|---|---|
| LMArena Text | 1372 | 1442 |
| LMArena Creative Writing | 1314 | 1410 |
| LMArena Multi-Turn | 1382 | 1447 |
Frequently asked questions
Is GPT-5.4 nano better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 41.9 on the Noometry Index. GPT-5.4 nano costs 3.1× 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-5.4 nano or Step 5 Preview?
GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is GPT-5.4 nano or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 43.6 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-5.4 nano and Step 5 Preview share?
17 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Step 5 Preview has 18.