Claude Sonnet 4.5
Best model in the world for agents, coding, and computer use
Claude Sonnet 4.5 is Anthropic's September 29, 2025 Sonnet model release: a hybrid reasoning large language model with strengths in coding, agentic tasks, computer use, reasoning, and math. Anthropic documentation lists the pinned Claude API ID as claude-sonnet-4-5-20250929 with alias claude-sonnet-4-5 and a 200k-token context window.
Pricing
Official site · Aug 10, 2026, 7:24 AM
Input / 1M
$3.00
Output / 1M
$15.00
Cached input / 1M
$0.30
Official table also lists 5-minute cache writes at $3.75/MTok and 1-hour cache writes at $6/MTok. MTok = million tokens.
Anthropic's official Claude Platform pricing page lists Claude Sonnet 4.5 at $3 per million base input tokens and $15 per million output tokens, with cache hits & refreshes at $0.30 per million tokens.
Model Intelligence
Context window
200,000 tokens
Arena ranking
30
Benchmarkable
Yes
Model level
release
Intelligence Index
29.9
Math Index
37
MMLU Pro
0.86
GPQA
0.73
HLE
0.07
LiveCodeBench
0.59
SciCode
0.43
AIME 2025
0.37
IFBench
0.43
LCR
0.52
TerminalBench Hard
0.29
TAU2
0.7
Recent stories
0 linked stories
No linked stories yet.