Claude Sonnet 5.5 · GDPval-AA(v2.1),Max effort配置,Elo rating 1844 Elo(具体成绩公布时间未确认)。见《Claude Sonnet 5.5 System Card》§8.14.3, PDF p.133。 查看成绩条件
Claude Sonnet 5.5 · GDPval-AA(v2.1),Xhigh effort配置,Elo rating 1725 Elo(具体成绩公布时间未确认)。见《Claude Sonnet 5.5 System Card》§8.14.3, PDF p.133。 查看成绩条件
Claude Sonnet 5.5 scores an Elo of 1844 at max effort, only 2 points below Claude Opus 5.5 at max effort (1846). At xhigh effort, Sonnet 5.5 scores 1725
Claude Sonnet 5.5 · GDPval-AA (v2.1), configuration Max effort, Elo rating 1844 Elo (exact score publication date unconfirmed). See “Claude Sonnet 5.5 System Card” at §8.14.3, PDF p.133. View score conditions
Claude Sonnet 5.5 · GDPval-AA (v2.1), configuration Xhigh effort, Elo rating 1725 Elo (exact score publication date unconfirmed). See “Claude Sonnet 5.5 System Card” at §8.14.3, PDF p.133. View score conditions
Claude Sonnet 5.5 scores an Elo of 1844 at max effort, only 2 points below Claude Opus 5.5 at max effort (1846). At xhigh effort, Sonnet 5.5 scores 1725
PDF cover dated 2026-09-28; 148 pages; collected 2026-11-10
Version published
Unconfirmed
Disclosure date
Unconfirmed · Launch or later disclosure unconfirmed
Collected
Historical basis
No reliable historical version evidence available
Citation text
Claude Sonnet 5.5 · GDPval-AA (v2.1), configuration Max effort, Elo rating 1844 Elo (exact score publication date unconfirmed); Claude Sonnet 5.5 · GDPval-AA (v2.1), configuration Xhigh effort, Elo rating 1725 Elo (exact score publication date unconfirmed). Official source: Claude Sonnet 5.5 System Card (Attachment, PDF cover dated 2026-09-28; 148 pages; collected 2026-11-10), §8.14.3, PDF p.133, https://www.anthropic.com/claude-sonnet-5-5-system-card#page=133. Benchlog https://benchlog.lekshome.cc/evidence/claude-sonnet-5-5/ev55-gdpval/?lang=en
Related scores and conditions
Claude Sonnet 5.5 · GDPval-AA scores and configurations
Metric and score
Configuration and evaluation conditions
Evidence relation
Elo rating1844 EloReported 1844 Elo · Unit Elo
Published Unconfirmed
Max effort
Settings
Max reasoning effort.
Unknowns
Adaptive thinking status is benchmark-specific.
Tools
Tools allowed
Reasoning budget
max
Framework
not recorded
Protocol
GDPval-AA v2.1, 220 tasks across 44 occupations and nine industries; independent Artificial Analysis evaluation.
Conditions
Shell and web browsing in agentic loop; pre-release deployment had a structured-output bug that may understate results and has since been fixed.