About
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,...
Related Models
Claude Sonnet 5 (Non-reasoning, High Effort)2026-06-30Claude Sonnet 5 (Adaptive Reasoning, High Effort)2026-06-30Claude Sonnet 5 (Adaptive Reasoning, Low Effort)2026-06-30Claude Sonnet 5 (Adaptive Reasoning, Medium Effort)2026-06-30Claude Sonnet 5 (Adaptive Reasoning, Xhigh Effort)2026-06-30Anthropic: Claude Fable 52026-06-09Anthropic: Claude Opus 4.82026-05-27Anthropic: Claude Opus 4.8 (Fast)2026-05-27
Pricing
Input
$2.00
per 1M tokens
Output
$10.00
per 1M tokens
Blended
$4.00
per 1M tokens
Cheaper than 16% of models. Median price is $0.56/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$4.00
Monthly
$120.00
vs. Similar Models
Anthropic: Claude Opus 4.7Q:+0.1
$10.00+150%
GPT-5.5 (high)Q:-0.3
$11.25+181%
OpenAI: GPT-5.5Q:+1.4
$11.25+181%
OpenAI: GPT-5.4Q:-2.0
$5.63+41%
Performance
88
tokens/sec
Faster than 49% of models
143.15
seconds
Faster than 0% of models
143.15
seconds
Faster than 0% of models
Market Median
91 tok/s
3% slower
Median TTFT
1.10s
12878% slower
Throughput/Dollar
22
tok/s per $/1M
Speed Comparison
OpenAI: GPT-4o (2024-05-13)
88 tok/s-0%
GPT-5.1 (Non-reasoning)
87 tok/s-1%
Ring-flash-2.0
86 tok/s-2%
Context Window
1.0M
tokens
Larger than 80% of models
Max Output
128K
tokens
13% of context
Benchmarks
MMLU-ProNot evaluated
GPQA Diamond
91.1%
HLE
39.6%
LiveCodeBenchNot evaluated
SciCode
53.6%
TerminalBench HardNot evaluated
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBenchNot evaluated
Long Context Recall
70.7%
Tau2Not evaluated
Market AverageTop Score