Claude Sonnet 5

ModelActive

by Anthropic · family “claude-sonnet”

Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,... (description from the OpenRouter listing)

in: textin: imagein: fileout: textReasoningTool useStructured output

Epoch Capabilities Index

156.3 #20 of 268

90% CI 153.4 – 159.2

Listing ↗

Released

30 Jun 2026

Context window

1M tokens

Max output

128K tokens

Input $ / 1M tokens

$2.00

Output $ / 1M tokens

$10.0

Cached input $ / 1M

$0.20

Each benchmark shown separately with its own source

Benchmark results (12)

All benchmarks

Benchmark	Domain	Score	vs best recorded	Setting	Run	Source
LMCA	agents	50.0%	76%	high	—	External ↗
DTBench	reasoning	92.5%	94%	max	—	External ↗
Mystery Game Puzzles	games	35.0% ±4.8	42%	max	28 Jul 2026	Epoch ↗
FrontierMath-Tiers-1-3-v2-Private	math	65.6% ±2.8	70%	max	30 Jun 2026	Eval log ↗
FrontierMath-Tier-4-v2-Private	math	29.3% ±7.2	30%	max	30 Jun 2026	Eval log ↗
DeepSWE	coding	53.8%	73%	max	—	External ↗
ProofBench	math	77.0%	77%	max	—	External ↗
Chess Puzzles	games	16.0% ±3.7	22%	max	30 Jun 2026	Eval log ↗
SimpleQA Verified	knowledge	33.7% ±1.5	45%	max	27 Aug 2026	Eval log ↗
WeirdML	coding	68.8% ±0.0	74%	high	—	External ↗
OTIS Mock AIME 2024-2025	math	80.0% ±6.0	80%	max	6 Aug 2026	Eval log ↗
GPQA diamond	science	80.3% ±2.8	84%	max	6 Aug 2026	Eval log ↗

Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.

API list price over time

Price history

$ per 1M tokens

4 observations (OpenRouter listing + Internet Archive snapshots). History accumulates with every ingest run; a single point means no change has been observed yet.