Model comparison
Claude Sonnet 5 vs GPT-5.6 Terra
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 54.6 on the Noometry Index.
Last verified . 46 shared benchmarks.
Summary
- They share 46 benchmarks with published results for both. Claude Sonnet 5 scores higher in 2 categories and GPT-5.6 Terra in 8 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Terra leads 81.6 to 66.2.
- The biggest single-benchmark swing is FrontierMath Tier 4: 29.3% for Claude Sonnet 5 and 70.7% for GPT-5.6 Terra.
- Claude Sonnet 5 is cheaper at $2 / $10 per million input/output tokens, against $2 / $12 for GPT-5.6 Terra.
- GPT-5.6 Terra accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Sonnet 5 | GPT-5.6 Terra | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 54.6 | 59.2 |
| Released | 2026-06-29 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $2 | $2 |
| Output $ / M tokens | $10 | $12 |
| Results tracked | 51 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Terra leads
Claude Sonnet 5: 55.5 (#26), GPT-5.6 Terra: 57.7 (#19)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| DeepSWE | 53.8% | 69.6% |
| FrontierCode | 42.7% | 41.3% |
| CursorBench | 34.1% | 41.3% |
| LMArena WebDev | 1541 | 1522 |
| SciCode | 54.3% | 55% |
| WeirdML | 68.8% | 78.3% |
| LMArena Coding | 1483 | 1484 |
| ALE-Bench | 1,463 | 1,951 |
| GSO | 37.3% | — |
Agentic & Tool Use Claude Sonnet 5 leads
Claude Sonnet 5: 42.8 (#18), GPT-5.6 Terra: 40.1 (#25)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| APEX-Agents | 54.5% | 58.2% |
| Vending-Bench 2 | 6,378 | 7,343 |
| BALROG | — | 53.2% |
| GBAEval | 65.3% | — |
| GDP.pdf | — | 24.7% |
| LMArena Search | 1194 | — |
Reasoning GPT-5.6 Terra leads
Claude Sonnet 5: 49.1 (#39), GPT-5.6 Terra: 60.7 (#21)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| SimpleBench | 60.6% | 48.9% |
| NYT Connections (extended) | 75.1% | 78.4% |
| CritPt | 16.9% | 30% |
| Chess Puzzles | 35% | 54% |
| LMArena Hard Prompts | 1461 | 1468 |
| Mystery Game Puzzles | 35% | 35% |
| DTBench | 92.5% | 93.3% |
| LMCA | 50% | 55% |
| Surface Evolver Bench | 60% | 83.8% |
| Epoch Capabilities Index | 156.21 | 159.62 |
| ARC-AGI-2 | — | 83.9% |
| Kagi LLM Benchmark | — | 51.3% |
| ARC-AGI-1 | — | 96.5% |
| Bench to the Future 3 | 0.14 | — |
| ForecastBench | 61.1 | — |
Math GPT-5.6 Terra leads
Claude Sonnet 5: 66.2 (#27), GPT-5.6 Terra: 81.6 (#12)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| FrontierMath (Tiers 1-3) | 65.6% | 86% |
| FrontierMath Tier 4 | 29.3% | 70.7% |
| OTIS Mock AIME 2024-2025 | 94.7% | 99.7% |
| ProofBench | 77% | 74% |
| LMArena Math | 1467 | 1466 |
Knowledge GPT-5.6 Terra leads
Claude Sonnet 5: 55.6 (#47), GPT-5.6 Terra: 61.2 (#30)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| GPQA Diamond | 90.5% | 93.3% |
| SimpleQA Verified | 33.7% | 43.2% |
| LMArena Expert | 1490 | 1492 |
Multimodal GPT-5.6 Terra leads
Claude Sonnet 5: 42.4 (#31), GPT-5.6 Terra: 47.3 (#11)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| LMArena Vision | 1274 | 1271 |
| Blueprint-Bench 2 | 24.9% | 30.8% |
| LMArena Document | 1466 | 1472 |
| Furniture Assembly | — | 54.2% |
Multilingual Too close to call
Claude Sonnet 5: 53.8 (#55), GPT-5.6 Terra: 54.4 (#44)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| LMArena Non-English | 1431 | 1439 |
| LMArena Chinese | 1477 | 1513 |
| LMArena French | 1460 | 1471 |
| LMArena German | 1440 | 1460 |
| LMArena Japanese | 1422 | 1457 |
| LMArena Korean | 1411 | 1425 |
| LMArena Russian | 1451 | 1450 |
| LMArena Spanish | 1437 | 1448 |
Instruction Following Too close to call
Claude Sonnet 5: 76.3 (#41), GPT-5.6 Terra: 76.4 (#40)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| LMArena Instruction Following | 1452 | 1454 |
Long Context Too close to call
Claude Sonnet 5: 44.8 (#55), GPT-5.6 Terra: 44.4 (#68)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| LMArena Longer Query | 1463 | 1451 |
Writing & Preference Too close to call
Claude Sonnet 5: 69.2 (#25), GPT-5.6 Terra: 70.2 (#23)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Terra |
|---|---|---|
| LMArena Text | 1442 | 1447 |
| LMArena Creative Writing | 1416 | 1410 |
| EQ-Bench Creative Writing | 1794 | 1855 |
| EQ-Bench 4 | 1236 | 1234 |
| LMArena Multi-Turn | 1454 | 1449 |
Frequently asked questions
Is Claude Sonnet 5 better than GPT-5.6 Terra?
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 54.6 on the Noometry Index.
Which is cheaper, Claude Sonnet 5 or GPT-5.6 Terra?
Claude Sonnet 5 is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; GPT-5.6 Terra lists at $2 and $12.
Is Claude Sonnet 5 or GPT-5.6 Terra better for coding?
GPT-5.6 Terra scores higher on coding benchmarks: 57.7 versus 55.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Terra does, with 1.05M tokens against 1M.
How many benchmarks do Claude Sonnet 5 and GPT-5.6 Terra share?
46 benchmarks have published results for both models. Claude Sonnet 5 has 51 scored results on Noometry and GPT-5.6 Terra has 52.