Model comparison
GPT-5.6 Terra vs Step 3
GPT-5.6 Terra is the stronger model overall, scoring 59.1 to 39.9 on the Noometry Index.
Last verified . 17 shared benchmarks.
StepFunSummary
- They share 17 benchmarks with published results for both. GPT-5.6 Terra scores higher in 9 categories and Step 3 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Terra leads 81.5 to 34.9.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 51.3% for GPT-5.6 Terra and 62.3% for Step 3.
- Step 3 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.6 Terra | Step 3 | |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 59.1 | 39.9 |
| Released | 2026-07-09 | — |
| Weights | Proprietary | Open |
| Context window | 1.05M | — |
| Max output | 128K | — |
| Input $ / M tokens | $2 | — |
| Output $ / M tokens | $12 | — |
| Results tracked | 52 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Terra leads
GPT-5.6 Terra: 57.1 (#16), Step 3: 39.5 (#144)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Coding | 1484 | 1367 |
| DeepSWE | 69.6% | — |
| FrontierCode | 41.3% | — |
| CursorBench | 41.3% | — |
| LMArena WebDev | 1520 | — |
| SciCode | 55% | — |
| WeirdML | 78.3% | — |
| ALE-Bench | 1,951 | — |
Agentic & Tool Use Not comparable
GPT-5.6 Terra: 40.2 (#23), Step 3: —
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| APEX-Agents | 58.2% | — |
| BALROG | 53.2% | — |
| GDP.pdf | 24.7% | — |
| Vending-Bench 2 | 7,343 | — |
Reasoning GPT-5.6 Terra leads
GPT-5.6 Terra: 60.6 (#19), Step 3: 27.6 (#102)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 51.3% | 62.3% |
| LMArena Hard Prompts | 1467 | 1355 |
| ARC-AGI-2 | 83.9% | — |
| SimpleBench | 48.9% | — |
| NYT Connections (extended) | 78.4% | — |
| ARC-AGI-1 | 96.5% | — |
| CritPt | 30% | — |
| Chess Puzzles | 54% | — |
| Mystery Game Puzzles | 35% | — |
| DTBench | 93.3% | — |
| LMCA | 55% | — |
| Surface Evolver Bench | 83.8% | — |
| Epoch Capabilities Index | 159.31 | — |
Math GPT-5.6 Terra leads
GPT-5.6 Terra: 81.5 (#10), Step 3: 34.9 (#148)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Math | 1463 | 1366 |
| FrontierMath (Tiers 1-3) | 86% | — |
| FrontierMath Tier 4 | 70.7% | — |
| OTIS Mock AIME 2024-2025 | 99.7% | — |
| ProofBench | 74% | — |
Knowledge GPT-5.6 Terra leads
GPT-5.6 Terra: 61.3 (#27), Step 3: 36.2 (#161)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Expert | 1492 | 1333 |
| GPQA Diamond | 93.3% | — |
| SimpleQA Verified | 43.2% | — |
Multimodal GPT-5.6 Terra leads
GPT-5.6 Terra: 47.5 (#10), Step 3: 34.9 (#81)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Vision | 1271 | 1176 |
| Blueprint-Bench 2 | 30.8% | — |
| Furniture Assembly | 54.2% | — |
| LMArena Document | 1472 | — |
Multilingual GPT-5.6 Terra leads
GPT-5.6 Terra: 54.5 (#42), Step 3: 46.5 (#154)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Non-English | 1439 | 1327 |
| LMArena Chinese | 1514 | 1396 |
| LMArena German | 1462 | 1371 |
| LMArena Korean | 1420 | 1271 |
| LMArena Russian | 1450 | 1331 |
| LMArena Spanish | 1455 | 1372 |
| LMArena French | 1476 | — |
| LMArena Japanese | 1446 | — |
Instruction Following GPT-5.6 Terra leads
GPT-5.6 Terra: 76.6 (#38), Step 3: 70.6 (#159)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1453 | 1332 |
Long Context GPT-5.6 Terra leads
GPT-5.6 Terra: 44.4 (#64), Step 3: 40.3 (#152)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Longer Query | 1452 | 1326 |
Writing & Preference GPT-5.6 Terra leads
GPT-5.6 Terra: 70.4 (#21), Step 3: 54.5 (#146)
| Benchmark | GPT-5.6 Terra | Step 3 |
|---|---|---|
| LMArena Text | 1446 | 1351 |
| LMArena Creative Writing | 1411 | 1322 |
| LMArena Multi-Turn | 1451 | 1340 |
| EQ-Bench Creative Writing | 1855 | — |
| EQ-Bench 4 | 1234 | — |
Frequently asked questions
Is GPT-5.6 Terra better than Step 3?
GPT-5.6 Terra is the stronger model overall, scoring 59.1 to 39.9 on the Noometry Index.
Is GPT-5.6 Terra or Step 3 better for coding?
GPT-5.6 Terra scores higher on coding benchmarks: 57.1 versus 39.5 in the Noometry coding category.
How many benchmarks do GPT-5.6 Terra and Step 3 share?
17 benchmarks have published results for both models. GPT-5.6 Terra has 52 scored results on Noometry and Step 3 has 17.