Claude Sonnet 5.5 vs GPT-6.1 Sol
| Benchmark | Sonnet 5.5 | GPT-6.1 Sol |
|---|---|---|
| Humanity's Last Exam | 45.8 | 49.9 |
| Arena Text · overall | 1,475 · xhigh | 1,484 · max |
| Arena Text · Chinese | 1,536.9 · xhigh | 1,494.6 · max |
| Arena Text · English | 1,484.5 · xhigh | 1,454.2 · max |
| Arena Text · French | 1,496.9 · xhigh | 1,470.1 · max |
| Arena Text · instruction following | 1,488.5 · xhigh | 1,468.6 · max |
| Arena Text · non-English | 1,454.2 · xhigh | 1,437.5 · max |
| Arena Text · Russian | 1,445.2 · xhigh | 1,457.9 · max |
| Arena Text · Spanish | 1,450.2 · xhigh | 1,440.4 · max |
| AA Omniscience Index | 21 | 40 |
| SimpleQA Verified v1 | 46.5 · max | 73.9 · max |
| CritPt (physics) | 24.6 | 27.7 |
| AA Intelligence Index 4.3.2 | 46.8 | 47.8 |
| AA Omniscience · Accuracy | 52 | 60.4 |
| AA Omniscience · Hallucination Rate | 64.6 | 51.6 |
| AA-LCR (long context) 1.1 | 78 | 83.3 |
| GDP.pdf · AA all-pass | 25.2 | 30 |
| MLCR-AA | 75 · max | 33.9 · max |
| Chess Puzzles | 59 · max | 68 · max |
| FrontierMath Erdos | 2.9 · max | 4.4 · max |
| FrontierMath Tier 4 (v2) | 80.5 · max | 100 · max |
| FrontierMath Tiers 1–3 (v2) | 88.8 · max | 93.7 · max |
| Mystery Game Puzzles | 65 · max | 80 · max |
| Vals MedCode | 52.9 | 48.8 |
| Vals MedScribe | 91.1 | 86.5 |
| Vals Legal Research Bench | 48.1 | 38.5 |
| Vals Science Index | 59.1 | 59.6 |
| ARC-AGI-1 · public evaluation | — | 97.9 |
| ARC-AGI-1 · semi-private | — | 95.5 |
| ARC-AGI-2 · public evaluation | — | 90.6 |
| ARC-AGI-2 · semi-private | — | 86.7 |
| ARC-AGI-3 · semi-private | — | 10.6 |
| MathArena · ArXivMath · August 2026 | — | 96.5 · max |
| MathArena · ArXivMath · June 2026 | — | 94.4 · max |
| MathArena · ArXivMath · May 2026 | — | 95 · max |
| MathArena · ArXivMath · September 2026 | — | 91.5 · max |
| MathArena · BrokenArXiv · August 2026 | — | 86.3 · max |
| MathArena · BrokenArXiv · June 2026 | — | 100 · max |
| MathArena · BrokenArXiv · May 2026 | — | 90 · max |
| MathArena · BrokenArXiv · September 2026 | — | 82.7 · max |
| MathArena · ArXivLean · September 2026 | — | 35.9 · max |
Specs
| Spec | Sonnet 5.5 | GPT-6.1 Sol |
|---|---|---|
| Input, per 1M | $2 | $2 |
| Output, per 1M | $10 | $10 |
| Cost per task? | $0.88 | $0.21 |
| Speed, tok/s | 106 | 45 |
| Context | 1M | 1.05M |
| Weights | Closed | Closed |
| Released | Sep 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | ✓ | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | — | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ✓ | ✓ |
| Run code | ✓ | ✓ |
| Use a computer | ✓ | ✓ |
What breaks when you switch
Remove unsupported parameters›
With reasoning enabled, remove temperature, top_p and top_logprobs, plus Chat Completions logprobs or Responses message.output_text.logprobs includes.
Responses API for tool calling · Chat Completions›
Use Responses for tools; Chat Completions supports this model without tool calling.
No none or minimal reasoning›
The none and minimal reasoning efforts are unsupported; choose low, medium, high, xhigh or max.
More than 272K input tokens›
Prompts above 272K input tokens charge 2× input and cache rates and 1.5× output rates for the entire request.
Cache writes billed at 1.25x›
Cache writes cost 1.25× the uncached input token rate.
Ultrafast prices 6x Standard›
Ultrafast mode costs 6× Standard token prices.
Context: 1.05M → 1M
No forced tool use›
Forced tool_choice any or tool returns HTTP 400; remove the forced choice.
Replace disabled thinking›
Disabled thinking returns HTTP 400; use between_tools for the lowest thinking setting.
No manual thinking budgets›
Manual thinking budgets and extra between_tools fields return HTTP 400; remove manual budgets and send between_tools without display, budget_tokens or block_binding.
Remove sampling parameters›
Non-default temperature, top_p or top_k returns HTTP 400; remove these parameters.
No assistant prefill›
A prefilled final assistant turn returns HTTP 400; remove that prefill.
Old computer tool rejected · Claude API / Google Cloud›
The computer_20251124 tool returns HTTP 400 on the Claude API and Google Cloud; use computer_toolset_20260801 there.