Claude 3.7 Sonnet (Thinking 16K): benchmark results

Provider: Anthropic. Released 2025-02-24. Access: API.

Aggregate Rating 1653 ± 28, rank #550 of 2287 rated models, from 14 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Generalization V1 (Lechmazur)1.73Avg Rank (lower is better)93.8
PlatinumBench (MIT)0.93Category-balanced mean error rate (%)90.9
Elimination Game (Lechmazur)5.28TrueSkill μ86.4
Confabulation Leaderboard (Lechmazur)7.92Confabulation rate % (lower is better)80.8
Arena-Hard Creative Writing63.9Win Rate (%)77.8
Arena-Hard v259.8Win Rate (%)74.1
MATH Level 586.25Accuracy (%)72.9
Step Game (Lechmazur)2.76TrueSkill μ63.5
NYT Connections Older Models19.7Score (%)59.9
Arena-Hard v2 (GPT-4.1 Judge)47Win Rate (%)59.3
Epoch AI - GPQA Diamond76.77Accuracy (%)51
PGG-Bench (Lechmazur)9.96TrueSkill μ50

Interactive version: theaggregate.ai/model?slug=claude-3-7-sonnet-thinking-16k · How It Works · Data refreshed daily, snapshot 2026-11-22.