Major Release Anthropic · Released
Claude Sonnet 4.6
What is Claude Sonnet 4.6?
Claude Sonnet 4.6 is a Major Release AI model from Anthropic, released on February 17, 2026. It has a 500K tokens context window, costs $3.00 per million input tokens, and scores 1468 on Chatbot Arena Elo. It is available via the company's API.
Near-Opus quality at a fraction of the cost, with Agent Teams orchestration.
Anthropic's latest Sonnet with Agent Teams capability and near-Opus performance at a fraction of the cost
Key features
- Agent Teams: orchestrate 2-16 Claude instances
- Near-Opus performance at 1/5th cost
- 80.8% SWE-bench Verified
- Fast mode research preview
How much does Claude Sonnet 4.6 cost?
| Metric | Price |
|---|---|
| input per mtok | $3.00 |
| output per mtok | $15.00 |
| cached input per mtok | $0.30 |
| cache read per mtok | $0.30 |
Pricing data verified .
What context window does Claude Sonnet 4.6 have?
- Context window
- 500K tokens
- Max output
- 64K
- Input modalities
- text, image, PDF
- Output modalities
- text
How does Claude Sonnet 4.6 score on benchmarks?
| Benchmark | Score |
|---|---|
| SWE-bench Verified | 80.8% |
| MMLU | 92.1% |
| HumanEval | 95.2% |
| GPQA Diamond | 79.7% |
| AIME 2025 | 88.5% |
| TAU-bench | 71.2% |
| HLE | 10.8% |
| SciCode | 44.1% |
| TAU2-bench | 78.9% |
| TerminalBench-Hard | 42.4% |
| IF-Bench | 42.4% |
| LiveCodeBench Reasoning | 58.7% |
| AA Intelligence Index | 42.6 |
| Chatbot Arena Elo | 1468 |
Benchmark scores verified .
What's new
- Agent Teams: orchestrate 2–16 Claude instances in parallel
- +8.5pt on SWE-bench Verified vs Sonnet 4
- 1/5 the cost of Opus 4.5 at ~95% of coding quality
- Fast mode research preview for lower-latency inference
How does Claude Sonnet 4.6 compare to Claude Sonnet 4?
- SWE-bench Verified: 72.3% → 80.8% (+8.5pt)
- Input pricing: $3.00 → $3.00
- Context window: 200K tokens → 500K tokens