Claude Sonnet 5
Anthropic · 6 thinking configurations
Intelligence38.2
Input / 1M$2.00
Output / 1M$10.00
Context1M
Tariff: Anthropic · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost5 measured configurations · 1 without both values
Selected: thinking max · Click a point to switch configuration
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,...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
38.2 pointsDeepSWE · pass@1Software-engineering tasks on mini-swe-agent v1.1. Pass@1 across evaluated runs, with measured cost per task. The 95% intervals may overlap.
53.8 % ±4.2GPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
91.1 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
41.3 %SciCodeCode generation for scientific research problems.
54.3 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
80.5 %MMMU-ProMultimodal understanding of academic problems that include images.
77.3 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
40.1 %Specifications from the catalog
Inputtext, image, file
Outputtext
Tool callingYes
Structured outputYes
Speed76.6 tok/s
First token202.74 s
Maximum output128K
Catalog listing date2026-06-30
DeepSWE v1.1 · 113 tasks · mini-swe-agent · cost $26.40 / task · 268 mean steps · 214.1K mean output tokens. Published by the source 9/22/2026.