Model comparison
Claude Opus 4.7 vs GPT-6 Astra
GPT-6 Astra is the stronger model overall, scoring 70.3 to 58.4 on the Noometry Index. Claude Opus 4.7 costs 2.0× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. Claude Opus 4.7 scores higher in 3 categories and GPT-6 Astra in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 53.7.
- The biggest single-benchmark swing is FrontierMath Tier 4: 31.7% for Claude Opus 4.7 and 97.6% for GPT-6 Astra.
- Claude Opus 4.7 is cheaper at $5 / $25 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Opus 4.7 | GPT-6 Astra | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 58.4 | 70.3 |
| Released | 2026-04-14 | 2026-09-03 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $5 | $10 |
| Output $ / M tokens | $25 | $50 |
| Results tracked | 66 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
Claude Opus 4.7: 59.9 (#10), GPT-6 Astra: 70.5 (#3)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| FrontierCode | 38.5% | 53.3% |
| LMArena WebDev | 1557 | 1792 |
| SciCode | 54.5% | 56.5% |
| WeirdML | 76.4% | 93.6% |
| LMArena Coding | 1519 | 1488 |
| MirrorCode | 31.1% | 46.7% |
| ALE-Bench | 1,323 | 2,951 |
| SWE-bench Verified | 83.5% | — |
| DeepSWE | — | 74.1% |
| FrontierSWE | — | 65.5% |
| GSO | 44.1% | — |
Agentic & Tool Use GPT-6 Astra leads
Claude Opus 4.7: 47.9 (#8), GPT-6 Astra: 52.3 (#3)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| APEX-Agents | 49.2% | 64.7% |
| Vending-Bench 2 | 10,937 | 15,515 |
| Terminal-Bench | 80.2% | — |
| OSWorld 2.0 | 18.2% | — |
| Remote Labor Index | — | 20.8% |
| τ²-bench Banking | 40.2% | — |
| PostTrainBench | 28.6% | — |
| BALROG | — | 68.3% |
| ExploitBench | 26.5% | — |
| GBAEval | 43.8% | — |
| GDP.pdf | 21% | — |
| LMArena Search | 1233 | — |
Reasoning GPT-6 Astra leads
Claude Opus 4.7: 53.7 (#24), GPT-6 Astra: 85.1 (#1)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| ARC-AGI-2 | 75.8% | 95% |
| NYT Connections (extended) | 39% | 98.1% |
| ARC-AGI-1 | 93.5% | 98.5% |
| CritPt | 12% | 31.7% |
| Chess Puzzles | 30% | 72% |
| EBR-Bench | 19% | 76.2% |
| LMArena Hard Prompts | 1507 | 1464 |
| Mystery Game Puzzles | 28% | 84% |
| DTBench | 94.7% | 97.3% |
| LMCA | 52.2% | 64.4% |
| Epoch Capabilities Index | 156.39 | 166.6 |
| SimpleBench | 61.7% | — |
| Kagi LLM Benchmark | 80.7% | — |
| Thematic Generalization | 72.8% | — |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 60.3 | — |
Math GPT-6 Astra leads
Claude Opus 4.7: 66.5 (#20), GPT-6 Astra: 93.4 (#1)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| FrontierMath (Tiers 1-3) | 70.2% | 93.7% |
| FrontierMath Tier 4 | 31.7% | 97.6% |
| OTIS Mock AIME 2024-2025 | 97.8% | 100% |
| ProofBench | 54% | 99% |
| LMArena Math | 1499 | 1463 |
| MathArena Final-Answer Competitions | 73.6% | — |
| FrontierMath (Feb 2025 set) | 43.8% | — |
| FrontierMath Erdős | — | 2.9% |
| FrontierMath Tier 4 (v1) | 22.9% | — |
Knowledge GPT-6 Astra leads
Claude Opus 4.7: 62.6 (#20), GPT-6 Astra: 75.3 (#1)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| GPQA Diamond | 90.2% | 95.8% |
| Humanity's Last Exam | 36.2% | 54.8% |
| SimpleQA Verified | 51.7% | 75.6% |
| Vectara Hallucination Rate | 12% | 8.7% |
| LMArena Expert | 1521 | 1486 |
Multimodal GPT-6 Astra leads
Claude Opus 4.7: 41.5 (#29), GPT-6 Astra: 55.1 (#3)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| LMArena Vision | 1317 | 1280 |
| Blueprint-Bench 2 | 24.5% | 49.7% |
| Furniture Assembly | 33.3% | 80% |
| LMArena Document | 1495 | 1468 |
Multilingual Claude Opus 4.7 leads
Claude Opus 4.7: 57.5 (#9), GPT-6 Astra: 54.0 (#53)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| LMArena Non-English | 1481 | 1431 |
| LMArena Chinese | 1530 | 1495 |
| LMArena French | 1505 | 1467 |
| LMArena Russian | 1494 | 1434 |
| LMArena Spanish | 1497 | 1412 |
| LMArena German | 1494 | — |
| LMArena Japanese | 1472 | — |
| LMArena Korean | 1466 | — |
Instruction Following Claude Opus 4.7 leads
Claude Opus 4.7: 78.6 (#9), GPT-6 Astra: 76.6 (#37)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| LMArena Instruction Following | 1498 | 1454 |
Long Context Claude Opus 4.7 leads
Claude Opus 4.7: 46.1 (#24), GPT-6 Astra: 44.7 (#57)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| LMArena Longer Query | 1505 | 1460 |
Writing & Preference Too close to call
Claude Opus 4.7: 75.3 (#8), GPT-6 Astra: 75.7 (#7)
| Benchmark | Claude Opus 4.7 | GPT-6 Astra |
|---|---|---|
| LMArena Text | 1491 | 1443 |
| LMArena Creative Writing | 1486 | 1423 |
| EQ-Bench Creative Writing | 1914 | 2173 |
| LMArena Multi-Turn | 1506 | 1457 |
| EQ-Bench 4 | 1311 | — |
Frequently asked questions
Is Claude Opus 4.7 better than GPT-6 Astra?
GPT-6 Astra is the stronger model overall, scoring 70.3 to 58.4 on the Noometry Index. Claude Opus 4.7 costs 2.0× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4.7 or GPT-6 Astra?
Claude Opus 4.7 is cheaper. It lists at $5 per million input tokens and $25 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is Claude Opus 4.7 or GPT-6 Astra better for coding?
GPT-6 Astra scores higher on coding benchmarks: 70.5 versus 59.9 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 1M.
How many benchmarks do Claude Opus 4.7 and GPT-6 Astra share?
45 benchmarks have published results for both models. Claude Opus 4.7 has 66 scored results on Noometry and GPT-6 Astra has 51.