Claude Sonnet 4.6
AnthropicProprietaryNon-reasoningVerifiedReleased Feb 1, 2026Input: text, image, file
Overall score
55.9
Rank 65 of 222 · 43.6–68.2
Context window
200K
200,000 tokens
Price per 1M tokens
$3.00 in · $15.00 out
✓ Verified on Anthropic's pricing page
Blended price
$6.00
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Knowledge#6077.9
- 2Coding#5565.6
- 3Reasoning62.8
- 4Multimodal#4658.7
- 5Agentic & tool use#5754.4
- 6Instruction following#9948.0
Benchmark results (22)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Agentic & tool use
| Gert Labs | 62.92 | #8 of 52 |
| MazeBench | 0 | #20 of 42 |
| Terminal-Bench 2.0 | 59.1 | #20 of 39 |
| Claw-Eval | 67.8 | #10 of 36 |
| OSWorld-Verified | 72.1 | #21 of 35 |
| JobBench | 36.9 | #18 of 34 |
| CyberGym | 65.2 | #21 of 25 |
| OSWorld 2.0 | 8.3 | #21 of 24 |
| ApprenticeBench | 2 | #18 of 18 |
Coding
| SWE-bench Verified | 79.6 | #20 of 98 |
| FrontierCode 1.1 Main | 24.3 | #16 of 16 |
| React Native Evals | 80.6 | #6 of 15 |
| SWE-Rebench | 60.7 | #5 of 13 |
Knowledge
| GPQA | 89.9 | #27 of 117 |
| HLE | 49.0 | #18 of 69 |
| MMLU-Pro | 79.2 | #38 of 67 |
| SuperGPQA | 95.0 | #1 of 20 |
Reasoning
| ARC-AGI-2 | 58.3 | #31 of 41 |
| ARC-AGI-1 | 86.00 | #29 of 36 |
Multimodal
| Design Arena Website | 1286 | #21 of 91 |
| CharXiv | 77.4 | #36 of 42 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| ✓ Official (Anthropic) checked Oct 17, 2026 · Claude Sonnet 4.6 | $3.00 | $15.00 | $0.30 |
| BenchLM list price | $3.00 | $15.00 | $0.30 |
| OpenRouter anthropic/claude-sonnet-4.6 | $3.00 | $15.00 | $0.30 |
Anthropic's Claude Sonnet 4.6 announcement says pricing remains the same as Sonnet 4.5 at $3 input / $15 output per million tokens.
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text | 1458 | #41 · 70.9K votes |
| WebDev (coding) | 1523 | #53 · 22.8K votes |
| Vision | 1283 | #33 · 27.4K votes |
| Documents | 1482 | #11 · 57.8K votes |
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Similar score
- Kimi K2.5 (Reasoning)56.2 · $1.20
- GLM-5.156.3 · $2.15
- Qwen3.5 397B55.4 · $1.35
- GPT-5.2-Codex55.4 · $4.81
- Qwen 3.6 Max (preview)55.1 · —