Model Terminal

Claude Sonnet 4

Claude Sonnet 4 is Anthropic's mid-tier frontier language model balancing speed, cost, and capability for coding, reasoning, writing, and agentic tool-use workflows. It supports both near-instant responses and an extended-thinking mode for harder tasks. Source: supabase.

Value score

Not scored

Context

200K

tokens

Max output

64K

tokens

Price

$3/1M in · $15/1M out

Capability radar

Peer value bars

Claude 4 Opus95
Claude 4 Sonnet91
Claude 3.7 Sonnet88
Claude 3.5 Haiku80

Identity

Developer: Anthropic
Openness: Closed
Modalities: —
Release: 2025-05-22
Knowledge cutoff: 2025-03-01
Deprecation: —
API docs: Provider docs

Benchmarks

GPQA75.4 · 2025-05-22 · source
MMLU86.5 · 2025-05-22 · source
SWE-bench72.7 · 2025-05-22 · source
LMArena Elo1283 · 2025-06-01 · source

Compare nearby

Claude 4 Opus Compare
Claude 4 Sonnet Compare
Claude 3.7 Sonnet Compare
Claude 3.5 Haiku Compare

Pricing

Input / 1M: $3
Output / 1M: $15
Speed: —