tensor
.news
Integrity Index
Models
Compare
Labs
Papers
Methodology
About
← Labs
Anthropic
21 models tracked · known for Claude 3 Opus, Claude 3 Sonnet, Claude 3.5 Haiku
21
Models tracked
21
In our data
99.72
Best score
3
Notable
Models
Claude Fable 5
independently reproduced
99.72
Claude Opus 4.8
independently reproduced
98.33
Claude Opus 4.7
independently reproduced
97.8
Claude Sonnet 4.5
independently reproduced
97.73
Claude Haiku 4.5
independently reproduced
96.36
Claude Opus 4.6
independently reproduced
94.44
Claude 3.7 Sonnet
independently reproduced
91.16
Claude 2
independently reproduced
87.5
Claude Instant
self-reported
86.7
Claude Sonnet 4.6
independently reproduced
86.5
Claude Opus 4.5
independently reproduced
86.1
Claude Opus 4.1
independently reproduced
85.38
Claude Opus 4
independently reproduced
85.05
Claude Sonnet 4
independently reproduced
84.37
Claude 3.5 Sonnet (October 2024)
independently reproduced
83.07
Claude 3.5 Sonnet
independently reproduced
82
Claude 3 Opus
independently reproduced
79.47
Claude 3.5 Haiku
independently reproduced
73.5
Claude 3 Sonnet
independently reproduced
67.87
Claude 3 Haiku
independently reproduced
65.07
Claude 2.1
independently reproduced
64.67
304 source records · epoch_benchmarks. Every score is evidence-linked.
Follow the record
Other labs
Qwen
Meta AI
OpenAI
DeepSeek
NVIDIA
Google DeepMind
EleutherAI
Zhipu AI
Explore the record
Benchmark Integrity Index
Model Claim Ledger
Head-to-head comparisons