Model
Claude Sonnet 4.6
Claude workhorse for coding agents, careful analysis, and production cost control
9 providersreleased knowledge cutoff 2025-08-31closed weights
$3
best input /M · Impossibl
$15
best output /M
1M
context window
64K
max output
Compare providers (sorted by listed input price)
| Provider | Input /M | Output /M | Cache read /M | Cache write /M | Context | Capabilities | Status | Updated |
|---|---|---|---|---|---|---|---|---|
| Impossibl anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsno structuredvision | Mar 13, 2026 | |
| CrossModel anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsstructuredvision | Mar 13, 2026 | |
| Anthropic claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsstructuredvision | Mar 13, 2026 | |
| LLM Gateway anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsstructuredvision | Mar 13, 2026 | |
| NEAR AI Cloud anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsno structuredvision | Mar 13, 2026 | |
| Merge Gateway anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsstructuredvision | Mar 13, 2026 | |
| Perplexity Agent anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | — | 200K | reasoningtoolsno structuredvision | Feb 17, 2026 | |
| Eden AI anthropic/claude-sonnet-4-6 | $3 | $15 | $0.3 | $3.75 | 1M | reasoningtoolsstructuredvision | Mar 13, 2026 | |
| AnyAPI anthropic/claude-sonnet-4-6 | — | — | — | — | 1M | reasoningtoolsno structuredvision | Mar 13, 2026 |
Price badge
https://models.sutraworks.ai/badge/anthropic/claude-sonnet-4-6.svgLive SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.
Benchmarks
- SWE-Atlas Codebase QnA(score)report ↗31.2
- SWE-Atlas Refactoring(score)report ↗32.21
- SWE-Atlas Test Writing(score)report ↗31.76
- Artificial Analysis Coding Agent Index ↗(average pass@1)report ↗49.4
- SWE-Atlas Codebase QnA(pass@1)report ↗70.3
- SWE-Bench Pro ↗(pass@1)report ↗14.9
- Terminal-Bench ↗(pass@1)report ↗63.1
- Terminal-Bench ↗(success rate)report ↗67
- Humanity's Last Exam ↗(accuracy)report ↗34.6
- Humanity's Last Exam ↗(accuracy)report ↗46.8
- OSWorld-Verified ↗(success rate)report ↗78.5
Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.