Claude Opus 4.6 vs Claude Opus 4.7
| Spec | Claude Opus 4.6 | Claude Opus 4.7 |
|---|---|---|
| Respan Index | 44.9 (#57) | 50.0 (#43) |
| Blended price per 1M | $10 | $10 |
| Context window | 1M | 1M |
| Input / output per 1M | $5 / $25 | $5 / $25 |
| Max output | 128K | 128K |
| Input | text, vision | text, vision |
| Released | 2026-02-05 | 2026-04-16 |
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 Opus 4.6 | Claude Opus 4.7 |
|---|---|---|
| Coding | 29.0 Claude Opus 4.6 | 32.2 Claude Opus 4.7 |
| Agents | 55.7 Claude Opus 4.6 | 61.1 Claude Opus 4.7 |
| Reasoning | 42.7 Claude Opus 4.6 | 53.6 Claude Opus 4.7 |
| Knowledge & language | 47.4 Claude Opus 4.6 | 49.9 Claude Opus 4.7 |
| Human preference | 55.1 Claude Opus 4.6 | 55.4 Claude Opus 4.7 |
Benchmarks both report
Claude Opus 4.7 leads on 25 of the 31 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Claude Opus 4.6 | Claude Opus 4.7 |
|---|---|---|
| LMArena Elo | 1503 | 1501 |
| LMArena Vision (LMArena) | 1298 | 1299 |
| LMArena WebDev (LMArena) | 1546 | 1557 |
| MirrorCode (Epoch AI) | 33.3% | 43.3% |
| Chess Puzzles (Epoch AI) | 21% | 41% |
| APEX-Agents (Mercor) | 46.3% | 49.2% |
| EBR-bench (Epoch AI) | 12.7% | 19% |
| FrontierMath Tier 4 v2 (Epoch AI) | 26.8% | 31.7% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 66% | 70.2% |
| Furniture Assembly (Epoch AI) | 28.3% | 33.3% |
| GPQA Diamond | 91.3% | 94.2% |
| LiveBench Agentic Coding (LiveBench) | 49% | 50.7% |
| LiveBench Coding (LiveBench) | 78.2% | 82.1% |
| LiveBench Data Analysis (LiveBench) | 69.9% | 78.3% |
| LiveBench Instruction Following (LiveBench) | 63.3% | 66.7% |
| LiveBench Language (LiveBench) | 83.3% | 77.9% |
| LiveBench Mathematics (LiveBench) | 89.3% | 92.9% |
| LiveBench Reasoning (LiveBench) | 88.7% | 87.2% |
| Mystery Game Puzzles (Epoch AI) | 25% | 28% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 94.4% | 97.8% |
| SAGE (Vals AI) | 51.6% | 56.1% |
| SimpleQA Verified | 47% | 51.7% |
| SWE-bench Verified (Epoch AI) | 78.7% | 83.5% |
| Vending-Bench 2 (Andon Labs) | 8017.59 | 10936.76 |
| WeirdML (Håvard Tveit Ihle) | 78% | 76.4% |
| AIME 2026 | 96.7% | 95.8% |
| tau2-bench Banking Knowledge (Sierra) | 27.3% | 40.2% |
| GSO Opt@1 (GSO) | 37.3% | 42.2% |
| SimpleBench (SimpleBench) | 67.6% | 61.7% |
| OSWorld-Verified | 72.7% | 82.8% |
| SWE-bench Verified | 80.8% | 87.6% |
FAQ
Is Claude Opus 4.6 better than Claude Opus 4.7?
Claude Opus 4.7 scores 50.0 on the Respan Index (#43), ahead of Claude Opus 4.6 at 44.9 (#57). Claude Opus 4.7 leads on 25 of the 31 benchmarks both report.
Which is cheaper, Claude Opus 4.6 or Claude Opus 4.7?
Claude Opus 4.6 lists at $5 input and $25 output per 1M tokens; Claude Opus 4.7 at $5 and $25. Both cost about the same per token at list price.
Which has the larger context window?
Both accept 1M tokens of context.