Model comparison
GPT-6 Luna vs Step 3
GPT-6 Luna is the stronger model overall, scoring 45.1 to 39.9 on the Noometry Index.
Last verified . 13 shared benchmarks.
StepFunSummary
- They share 13 benchmarks with published results for both. GPT-6 Luna scores higher in 9 categories and Step 3 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-6 Luna leads 54.7 to 39.5.
- Step 3 has downloadable open weights; the other is API-only.
Side by side
| GPT-6 Luna | Step 3 | |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 45.1 | 39.9 |
| Released | 2026-09-22 | — |
| Weights | Proprietary | Open |
| Context window | 1.05M | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.10 | — |
| Output $ / M tokens | $0.50 | — |
| Results tracked | 18 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
GPT-6 Luna: 54.7 (#24), Step 3: 39.5 (#144)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Coding | 1448 | 1367 |
| DeepSWE | 66.6% | — |
| FrontierCode | 42.4% | — |
| LMArena WebDev | 1583 | — |
Reasoning GPT-6 Luna leads
GPT-6 Luna: 31.7 (#68), Step 3: 27.6 (#102)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Hard Prompts | 1415 | 1355 |
| Kagi LLM Benchmark | — | 62.3% |
| NYT Connections (extended) | 68.7% | — |
Math Too close to call
GPT-6 Luna: 35.6 (#140), Step 3: 34.9 (#148)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Math | 1396 | 1366 |
Knowledge GPT-6 Luna leads
GPT-6 Luna: 40.1 (#110), Step 3: 36.2 (#161)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Expert | 1458 | 1333 |
Multimodal GPT-6 Luna leads
GPT-6 Luna: 42.0 (#27), Step 3: 34.9 (#81)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Vision | — | 1176 |
| Furniture Assembly | 44.2% | — |
Multilingual GPT-6 Luna leads
GPT-6 Luna: 51.0 (#111), Step 3: 46.5 (#154)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Non-English | 1390 | 1327 |
| LMArena Chinese | 1425 | 1396 |
| LMArena Russian | 1392 | 1331 |
| LMArena Spanish | 1416 | 1372 |
| LMArena German | — | 1371 |
| LMArena Korean | — | 1271 |
Instruction Following GPT-6 Luna leads
GPT-6 Luna: 74.7 (#85), Step 3: 70.6 (#159)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1414 | 1332 |
Long Context GPT-6 Luna leads
GPT-6 Luna: 43.3 (#95), Step 3: 40.3 (#152)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Longer Query | 1417 | 1326 |
Writing & Preference GPT-6 Luna leads
GPT-6 Luna: 58.4 (#115), Step 3: 54.5 (#146)
| Benchmark | GPT-6 Luna | Step 3 |
|---|---|---|
| LMArena Text | 1392 | 1351 |
| LMArena Creative Writing | 1357 | 1322 |
| LMArena Multi-Turn | 1400 | 1340 |
Frequently asked questions
Is GPT-6 Luna better than Step 3?
GPT-6 Luna is the stronger model overall, scoring 45.1 to 39.9 on the Noometry Index.
Is GPT-6 Luna or Step 3 better for coding?
GPT-6 Luna scores higher on coding benchmarks: 54.7 versus 39.5 in the Noometry coding category.
How many benchmarks do GPT-6 Luna and Step 3 share?
13 benchmarks have published results for both models. GPT-6 Luna has 18 scored results on Noometry and Step 3 has 17.