Claude Opus 5.5 vs GPT-6 Astra
Published test results, prices, limits and features, side by side. Share the link to show someone exactly this comparison.
Claude Opus 5.5Anthropic
GPT-6 AstraOpenAI
At a glance
Capability index
- Claude Opus 5.5167.3
- GPT-6 Astra166.4
Claude Opus 5.5 is 0.9 points higher.
Price per 1M tokens
- Claude Opus 5.5$8.00
- GPT-6 Astra$20.00
Claude Opus 5.5 is 2.5 times cheaper.
Context window
- Claude Opus 5.51M tokens
- GPT-6 Astra1.1M tokens
GPT-6 Astra takes in 5% more.
Released
- Claude Opus 5.5Sep 22, 2026
- GPT-6 AstraSep 4, 2026
Claude Opus 5.5 is 18 days newer.
Where each scores higher
On the 20 tests both have taken, each on its own scale. A gap of a point or two can sit within a test’s margin of error.
Claude Opus 5.5
- LiveBench overall+1.1 points
- Arena text+27 points
- LiveBench coding+8.9 points
- Arena WebDev+28 points
- LiveBench mathematics+0.3 points
- Arena math+20 points
- LiveBench agentic coding+14.4 points
- Arena creative writing+67 points
- Arena instruction following+44 points
- Arena hard prompts+32 points
GPT-6 Astra
- Arena coding+5 points
- FrontierMath Tiers 1–3+2.5 points
- FrontierMath Tier 4+2.6 points
- SimpleQA Verified+3.4 points
- LiveBench language+3.2 points
- LiveBench instructions+8.5 points
- LiveBench data analysis+2.7 points
- LiveBench reasoning+0.5 points
- GPQA Diamond+5.2 points
Level on Mock AIME 2024–2025.
3 more tests have results for only one of them; the table below lists every result.
Everything side by side
| Measure | Claude Opus 5.5Anthropic | GPT-6 AstraOpenAI |
|---|---|---|
| Model | ||
| Maker | Anthropic | OpenAI |
| Sold here by | Anthropic | OpenAI |
| API model ID | claude-opus-5-5 | gpt-6-astra |
| Released | Sep 22, 2026 | Sep 4, 2026 |
| Knowledge cutoff | Jun 2026 | Apr 30, 2026 |
| Weights | Closed: hosted access only | Closed: hosted access only |
| Developer’s country | United States | United States |
| Hosts selling it | 35including Anthropic directly | 29including OpenAI directly |
| Training compute | Not published | 1.0 × 10²⁷ FLOPEpoch AI estimate |
| Test results | ||
| Capability index | 167.3 (best of these models)#1 of 274 | 166.4#2 of 274 |
| LiveBench overall | 83.2% (best of these models)#2 of 66 | 82.2%#4 of 66 |
| Arena text | 1504 (best of these models)#4 of 413 | 1477#29 of 413 |
| LiveBench coding | 89.3% (best of these models)#2 of 66 | 80.4%#20 of 66 |
| Arena WebDev | 1815 (best of these models)#1 of 138 | 1788#2 of 138 |
| Arena coding | 1539#12 of 408 | 1543 (best of these models)#7 of 408 |
| LiveBench mathematics | 97.1% (best of these models)#1 of 66 | 96.8%#4 of 66 |
| Mock AIME 2024–2025 | 100.0%#1 of 297 | 100.0%#1 of 297 |
| FrontierMath Tiers 1–3 | 91.2%#3 of 114 | 93.7% (best of these models)#1 of 114 |
| FrontierMath Tier 4 | 95.0%#6 of 70 | 97.6% (best of these models)#2 of 70 |
| Arena math | 1511 (best of these models)#8 of 396 | 1490#27 of 396 |
| SimpleQA Verified | 72.2%#4 of 86 | 75.6% (best of these models)#1 of 86 |
| LiveBench agentic coding | 71.7% (best of these models)#2 of 66 | 57.3%#21 of 66 |
| CursorBench | 57.8%#1 of 62 | Not tested |
| DeepSWE | Not tested | 74.1%#1 of 69 |
| LiveBench language | 86.3%#10 of 66 | 89.4% (best of these models)#4 of 66 |
| LiveBench instructions | 67.0%#39 of 66 | 75.6% (best of these models)#8 of 66 |
| Arena creative writing | 1516 (best of these models)#2 of 411 | 1449#39 of 411 |
| Arena instruction following | 1517 (best of these models)#2 of 413 | 1474#30 of 413 |
| LiveBench data analysis | 80.3%#7 of 66 | 83.0% (best of these models)#1 of 66 |
| LiveBench reasoning | 92.2%#3 of 66 | 92.7% (best of these models)#1 of 66 |
| GPQA Diamond | 90.6%#40 of 319 | 95.8% (best of these models)#1 of 319 |
| Arena hard prompts | 1534 (best of these models)#2 of 413 | 1502#30 of 413 |
| Chess puzzles | Not tested | 72.0%#1 of 227 |
| Price per million tokens | ||
| Input | $4.00 (best of these models) | $10.00 |
| Output | $20.00 (best of these models) | $50.00 |
| Blended, 3 input to 1 output | $8.00 (best of these models) | $20.00 |
| Cached input | $0.20 (best of these models) | $1.00 |
| Writing to the cache | $5.00 (best of these models) | $12.50 |
| Long requests | No separate price reported | $20.00 in, $75.00 outabove 272K tokens |
| The maker’s own price | This listing | This listing |
| Limits | ||
| Context window | 1M tokens | 1.1M tokens (best of these models) |
| Max input | 1M tokens (best of these models) | 922K tokens |
| Max output | 128K tokens | 128K tokens |
| Features | ||
| Reads | images, PDFs and text | images, PDFs and text |
| Produces | text | text |
| Reasoning | Yeseffort low, medium, high, xhigh, max; can be switched off | Yeseffort low, medium, high, xhigh, max |
| Tool calling | Yes | Yes |
| Structured output | Yes, such as JSON | Yes, such as JSON |
| File attachments | Yes | Yes |
| Temperature setting | Fixed by the provider | Fixed by the provider |
Bold marks the better value in each row. Prices are each model’s own list price, or the middle price across hosts when the maker does not sell it, blended as three input tokens for every output token. Test results as published by LiveBench (Apache-2.0), Epoch AI (CC BY 4.0), Datacurve DeepSWE, compiled by Epoch AI (CC BY 4.0), Cursor CursorBench, compiled by Epoch AI (CC BY 4.0), Arena (CC BY 4.0); model details and prices from models.dev (MIT). Confirm pricing with the provider before use.