Claude Sonnet 4.5 vs Claude Sonnet 4.6
| Spec | Claude Sonnet 4.5 | Claude Sonnet 4.6 |
|---|---|---|
| Respan Index | 31.3 (#85) | 37.3 (#73) |
| Blended price per 1M | $6 | $6 |
| Context window | 200K | 1M |
| Input / output per 1M | $3 / $15 | $3 / $15 |
| Max output | 64K | 128K |
| Input | text, vision | text, vision |
| Released | 2025-09-29 | 2026-02-17 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Claude Sonnet 4.5 | Claude Sonnet 4.6 |
|---|---|---|
| Coding | 13.0 Claude Sonnet 4.5 | 14.9 Claude Sonnet 4.6 |
| Agents | 41.5 Claude Sonnet 4.5 | 50.9 Claude Sonnet 4.6 |
| Reasoning | 27.5 Claude Sonnet 4.5 | 36.5 Claude Sonnet 4.6 |
| Knowledge & language | 31.2 Claude Sonnet 4.5 | 35.7 Claude Sonnet 4.6 |
| Human preference | 51.0 Claude Sonnet 4.5 | 54.5 Claude Sonnet 4.6 |
Benchmarks both report
Claude Sonnet 4.6 leads on 22 of the 25 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Claude Sonnet 4.5 | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Elo | 1456 | 1472 |
| LMArena WebDev (LMArena) | 1394 | 1522 |
| Chess Puzzles (Epoch AI) | 10% | 23% |
| GPQA Diamond | 83.4% | 89.9% |
| Mystery Game Puzzles (Epoch AI) | 17% | 16% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 77.8% | 85.8% |
| SAGE (Vals AI) | 36.1% | 46.6% |
| SimpleQA Verified | 30.7% | 35.5% |
| SWE-bench Verified (Epoch AI) | 71.3% | 75.2% |
| Vending-Bench 2 (Andon Labs) | 3838.74 | 7204.14 |
| WeirdML (Håvard Tveit Ihle) | 47.7% | 66.1% |
| ARC-AGI-2 (Verified) | 13.6% | 58.3% |
| GDPval-AA | 1276 | 1633 |
| Humanity's Last Exam (no tools) | 17.7% | 33.2% |
| Humanity's Last Exam (with tools) | 33.6% | 49% |
| MCP Atlas | 43.8% | 61.3% |
| MMMLU | 89.5% | 89.3% |
| MMMU-Pro (no tools) | 63.4% | 74.5% |
| MMMU-Pro (with tools) | 68.9% | 75.6% |
| OSWorld-Verified | 61.4% | 72.5% |
| SWE-bench Verified | 77.2% | 79.6% |
| Tau2-bench Retail | 86.2% | 91.7% |
| Tau2-bench Telecom | 98% | 97.9% |
| Terminal-Bench 2.0 | 51% | 59.1% |
| OSWorld-Verified (XLANG) | 62.9% | 72.1% |
FAQ
Is Claude Sonnet 4.5 better than Claude Sonnet 4.6?
Claude Sonnet 4.6 scores 37.3 on the Respan Index (#73), ahead of Claude Sonnet 4.5 at 31.3 (#85). Claude Sonnet 4.6 leads on 22 of the 25 benchmarks both report.
Which is cheaper, Claude Sonnet 4.5 or Claude Sonnet 4.6?
Claude Sonnet 4.5 lists at $3 input and $15 output per 1M tokens; Claude Sonnet 4.6 at $3 and $15. Both cost about the same per token at list price.
Which has the larger context window?
Claude Sonnet 4.5 accepts 200K tokens and Claude Sonnet 4.6 1M tokens.