Claude Sonnet 4.6 vs GPT-5.4
Head to head on 28 shared benchmarks, with price, context window, and release dates.
Benchmark scores
46evals · sort by any column, ⤢ to expand fullscreen
| Model | |||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Sonnet 4.6 Anthropic | 47.6% | 1.35 | 65.3% | 78.0% | 1207elo | 1805elo | 79.9 | 32.4% | 79.7% | 11.2% | 42.4% | 92.2% | 78.4% | 80.0% | 76.1% | 63.9% | 77.7% | 86.5% | 86.4% | 67.7% | 45.0% | 46.6% | 44.1% | 77.1% | 42.4% | 76.3 | 51.5% | 78.9% | 87.1% | 47.2% | - | - | 51.0% | - | - | - | - | - | 89.3 | - | - | 72.1% | - | - | 50.6% | 60.6% | 62.8% |
| GPT-5.4 OpenAI | 13.1% | 1.47 | 65.3% | 81.0% | 1272elo | 1836elo | 78.3 | 47.6% | 74.8% | 11.3% | 48.4% | 78.5% | 82.0% | 77.5% | 79.3% | 70.2% | 82.6% | 94.1% | 88.1% | 68.3% | 56.0% | 43.3% | 47.1% | 74.0% | 37.9% | 80 | 67.4% | 36.0% | - | - | 34.1% | 9.0% | - | 87.2% | 92063 | 82.8% | 41.3% | 77.5% | - | 86 | 52.2% | - | -0.61 | 40.8 | - | - | 60.3% |
2 / 2 models
| Attribute | Claude Sonnet 4.6 Anthropic | GPT-5.4 OpenAI |
|---|---|---|
| Description | Claude Sonnet 4.6 is an AI model from Anthropic. | GPT-5.4 is an AI model from OpenAI. |
| Family | Claude | GPT |
| Params | - | - |
| Open weights | closed | closed |
| License | proprietary | proprietary |
| Released | 17 Feb 2026 | 5 Mar 2026 |
| Context | 1,000,000 | 1,050,000 |
| Input $/1M | $3.00 | $2.50 |
| Output $/1M | $15.00 | $15.00 |
| Cache read $/1M | $0.300 | $0.250 |
| Cache write $/1M | $3.750 | - |
| Speed | 45 tok/s | 0 tok/s |
| Latency | 1.12 s | 0.00 s |
| Modalities | text, image, file | text, image, file |
| Providers | - | - |
| Publisher | Anthropic | OpenAI |
| Reported scores | GPQA Diamond79.7% (AA) Humanity's Last Exam (HLE)11.2% (AA) IFBench42.4% (AA) SciCode44.1% (AA) τ²-bench (Tau²-bench)78.9% (AA) | GPQA Diamond74.8% (AA) SciCode47.1% (AA) τ²-bench (Tau²-bench)36.0% (AA) Terminal-Bench (Hard)37.9% (AA) Vibe Code Bench v1.167.4% |
FAQ
- Is Claude Sonnet 4.6 or GPT-5.4 better?
- Across the 28 benchmarks both models report on Sophon, GPT-5.4 leads on 18 of them. Which one is "better" depends on the benchmark - the table above shows every shared score.
- How do Claude Sonnet 4.6 and GPT-5.4 compare on GPQA Diamond?
- Claude Sonnet 4.6 scores 79.7% and GPT-5.4 scores 74.8% on GPQA Diamond.
- When were Claude Sonnet 4.6 and GPT-5.4 released?
- Claude Sonnet 4.6 was released 17 Feb 2026 by Anthropic. GPT-5.4 was released 5 Mar 2026 by OpenAI.
Comparing something else? Build your own side-by-side.