Claude 3.5 Sonnet
Anthropic's previous best-in-class workhorse — exceptional coding, reasoning, and vision at an accessible price.
Pricing
Input$3 / 1M
Cached inputAvailable
Output$15 / 1M
Cost bandBalanced
Free tierClaude.ai free tier; API is paid.
Limits
Context window200k
Max output8,192
Knowledge cutoffApr 1, 2024
DeploymentClaude Platform, Bedrock, Vertex AI
OpenAI-compatible APINo
ModalitiesText, Image input
Where it fits
Best forTeams running on Claude 3.5 still; coding, data extraction, document analysis.
Business fitConsider migrating to Claude Sonnet 4.6 for better context and agent performance.
Capabilities
Tool useVisionComputer use beta
Benchmark scores
Composite83 / 100
Coding88 / 100 (SWE-bench Verified)
Reasoning82 / 100 (GPQA Diamond)
Long context84 / 100 (RULER 200K)
Vision78 / 100 (MMMU)
Instruction following85 / 100 (IFEval)
Editorial aggregate of published benchmark results as of Nov 1, 2024. These are not independent measurements run by this site.
Change history
- Oct 28, 2025DeprecatedClaude 3.5 Sonnet retired Retired from the Claude API on Oct 28, 2025. Anthropic's recommended replacement: Claude Sonnet 4.6. Anthropic model deprecations
Data
Last checkedJan 10, 2026
Verify with provider documentation before committing spend. Report an error and it gets fixed with a logged correction.