LLM ReferenceLLM Reference

Claude Sonnet 4

claude-sonnet-4

DeprecatedResearched 3d ago

Last refreshed 2026-05-15. Next refresh: weekly.

ProprietaryRAGLong contextVisionClassificationJSON / Tool use

Claude Sonnet 4 is a legacy integration reference; evaluate Claude Sonnet 4.6 before starting new work.

Decision context: Vision task fit, 3 tracked provider routes, and research from 2026-05-15.

Use it for

  • Teams maintaining an existing integration
  • Workloads that can use a 1M context window
  • Buyers comparing 3 tracked provider routes

Do not use it for

  • New production launches
  • Vision or document-understanding workloads

Cheapest output

$15.00

Anthropic per 1M tokens

Provider routes

3

Tracked API hosts

Quality / dollar

Unknown

No output-token price in the ladder

Freshness

2026-05-15

Researched 3d ago

fresh

This API model is marked deprecated. Use Claude Sonnet 4.6 as the replacement candidate before sending new production traffic.

Top use-case fit

RAG

Included by capability and metadata signals in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Vision

1 relevant benchmark in the decision map.

Provider price ladder

ProviderInput / 1MOutput / 1MRoute
Anthropic$3.00$15.00
Serverless
OpenRouter$3.00$15.00
Serverless
AWS Bedrock$8.00$40.00
Serverless

Benchmark peer barsfor Vision

Migration checks

No linked migration route is available for this model yet.

About

Claude Sonnet 4 API snapshot with expanded 1M-token context. Anthropic lists claude-sonnet-4-20250514 as deprecated April 14, 2026 with Claude API retirement scheduled for June 15, 2026; use Claude Sonnet 4.6 for current integrations.

Claude Sonnet 4 has a 1M-token context window.

Claude Sonnet 4 input tokens at $3/1M, output at $15/1M.

Capabilities

Structured Outputs

Benchmark Scores(1)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
Massive Multi-discipline Multimodal Understanding74.4https://mmmu-benchmark.github.io/

Specifications

FamilyClaude 4
Released2025-08-01
Context1M

Created by

Developing safe and ethical AI systems.

San Francisco, California, United States
Founded 2021
Website