Claude Opus 4.6vsOpus 4.5
Across 17 shared benchmarks, Claude Opus 4.6 leads overall: Claude Opus 4.6 wins 14, Opus 4.5 wins 3, with 0 ties and an average score difference of +8.99.
Claude Opus 4.6
Anthropic · 2026-02-05 · Reasoning model
Opus 4.5
Anthropic · 2025-11-25 · Reasoning model
Claude Opus 4.614 wins(82%)(18%)3 winsOpus 4.5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 17 shared benchmarks.
General Knowledge
Claude Opus 4.6 5/5| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| ARC-AGI-2 | 66.3016 / 61Extended (no tools) | 37.6028 / 61Extended (no tools) | +28.70 |
| ARC-AGI | 9212 / 67Extended (no tools) | 8023 / 67Extended (no tools) | +12 |
| HLE | 5316 / 170Extended (with tools, internet) | 43.2047 / 170Extended (with tools) | +9.80 |
| GPQA Diamond | 91.3115 / 187Extended (no tools) | 8742 / 187Extended (no tools) | +4.31 |
| LiveBench | 76.338 / 115Thinking High (No Tools) | 75.9611 / 115Thinking (No Tools, 64K Budget) | +0.37 |
Agent Level Benchmark
Claude Opus 4.6 2/2| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| τ²-Bench | 91.891 / 43Extended (with tools) | 81.9913 / 43Extended (with tools) | +9.90 |
| τ²-Bench - Telecom | 99.252 / 35Extended (with tools) | 90.7021 / 35Extended (with tools) | +8.55 |
AI Agent - Tool Usage
Claude Opus 4.6 2/2| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| MCP-Atlas | 76.8010 / 27Deep Thinking (With Tools) | 69.8016 / 27Thinking High (With Tools) | +7 |
| Terminal Bench 2.0 | 65.4011 / 47Extended (with tools) | 59.3020 / 47Extended (with tools) | +6.10 |
Coding and Software Engineer
Opus 4.5 2/2| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| LiveCodeBench | 7638 / 123Extended (no tools) | 8712 / 123Extended (with tools) | -11 |
| SWE-bench Verified | 80.849 / 111Extended (with tools) | 80.908 / 111Extended (with tools) | -0.06 |
Math and Reasoning
Claude Opus 4.6 2/2| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| FrontierMath | 40.707 / 60最高（无工具） | 20.7017 / 60Extended (no tools) | +20 |
| FrontierMath - Tier 4 | 22.9012 / 80最高（无工具） | 4.2040 / 80Normal (No Tools) | +18.70 |
Claw-style Agent Evaluation
Claude Opus 4.6 1/1| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| Pinch Bench | 87.407 / 37Thinking (With Tools) | 87.208 / 37Extended (with tools) | +0.20 |
Commonsense Reasoning
Claude Opus 4.6 1/1| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| Simple Bench | 67.608 / 63Normal (No Tools) | 6212 / 63Extended (no tools) | +5.60 |
Instruction Following
Claude Opus 4.6 1/1| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| IF Bench | 941 / 30Extended (no tools) | 5821 / 30Extended (with tools) | +36 |
Multimodal Understanding
Opus 4.5 1/1| Benchmark | Claude Opus 4.6 | Opus 4.5 | Diff |
|---|---|---|---|
| MMMU | 77.3016 / 29Extended (with tools) | 80.7011 / 29Extended (no tools) | -3.40 |
Specs
| Field | Claude Opus 4.6 | Opus 4.5 |
|---|---|---|
| Publisher | Anthropic | Anthropic |
| Release date | 2026-02-05 | 2025-11-25 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 1000K | 200K |
| Max output | 64K | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Claude Opus 4.6 | Opus 4.5 |
|---|---|---|
| Text input | $0.5 / 1M tokens | $5 / 1M tokens |
| Text output | $25 / 1M tokens | $25 / 1M tokens |
| Cache read | $0.5 / 1M tokens | $0.5 / 1M tokens |
| Cache write | $10 / 1M tokens | $6.25 / 1M tokens |
Summary
- Claude Opus 4.6leads in:General Knowledge (5/5), Agent Level Benchmark (2/2), AI Agent - Tool Usage (2/2), Math and Reasoning (2/2), Claw-style Agent Evaluation (1/1), Commonsense Reasoning (1/1), Instruction Following (1/1)
- Opus 4.5leads in:Coding and Software Engineer (2/2), Multimodal Understanding (1/1)
On average across the 17 shared benchmarks, Claude Opus 4.6 scores 8.99 higher.
Largest single-benchmark gap: IF Bench — Claude Opus 4.6 94 vs Opus 4.5 58 (+36).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.