anthropic

Claude Sonnet 4.6

claude-sonnet

Claude workhorse for coding agents, careful analysis, and production cost control

ReasoningTool CallingAttachments

Context Window

1M

Max Output

64K

Temperature

Yes

Open Weights

No

Knowledge Cutoff

2025-08-31

Released

2026-02-17

Last Updated

2026-03-13

Modalities

Input:TextImagePDF
→
Output:Text

Benchmarks

BenchmarkScoreMetricHarness
SWE-Atlas Codebase QnA31.2scoreClaude Code
SWE-Atlas Refactoring32.21scoreClaude Code
SWE-Atlas Test Writing31.76scoreClaude Code
Artificial Analysis Coding Agent Index49.4average pass@1Claude Code
SWE-Atlas Codebase QnA70.3pass@1Claude Code
SWE-Bench Pro14.9pass@1Claude Code
Terminal-Bench63.1pass@1Claude Code

Available Providers (6)

ProviderInput /1MOutput /1MCache Read /1MCache Write /1MReasoningStatus
RequestyRequestyRp 54.102Rp 270.506Rp 5.411Rp 67.627effort, budget_tokens—
AnyAPIAnyAPI————toggle, effort, budget_tokens—
Merge GatewayMerge GatewayRp 54.102Rp 270.506Rp 5.411Rp 67.627toggle, effort, budget_tokens—
NEAR AI CloudNEAR AI CloudRp 54.102Rp 270.506Rp 5.411Rp 67.627toggle, budget_tokens—
Cloudflare AI GatewayCloudflare AI GatewayRp 54.102Rp 270.506Rp 5.411Rp 67.627effort—
Perplexity AgentPerplexity AgentRp 54.102Rp 270.506Rp 5.411—effort—
← Kembali ke Katalog