tensor
.news
Integrity Index
Models
Compare
Labs
Papers
Methodology
About
← Labs
OpenAI
40 models tracked · known for gpt2, gpt-oss-20b, gpt-oss-120b
40
Models tracked
40
In our data
100
Best score
3
Notable
Models
GPT-5.5
independently reproduced
100
GPT-5.5 Pro
independently reproduced
100
GPT-5
independently reproduced
98.13
GPT-5 mini
independently reproduced
97.85
o4-mini
independently reproduced
97.83
o3
independently reproduced
97.77
o3-pro
unverified
97.2
o3-mini
independently reproduced
96.49
GPT-5.2
independently reproduced
96.11
GPT-5.4
independently reproduced
95.3
GPT-5 nano
independently reproduced
95.24
o1
independently reproduced
94.71
GPT-5.4 Pro
independently reproduced
94.5
GPT-4 (Mar 2023)
independently reproduced
93.73
GPT-4o mini
independently reproduced
91.3
GPT-5.2 Pro
independently reproduced
90.5
GPT-4 (Jun 2023)
independently reproduced
89.99
o1-mini
independently reproduced
89.18
gpt-oss-120b
independently reproduced
88.88
GPT-5.1
independently reproduced
88.6
GPT-5.4 Nano
independently reproduced
87.77
GPT-4.1 mini
independently reproduced
87.29
GPT-5.4 Mini
independently reproduced
87.21
GPT-3.5 Turbo (Nov 2023)
independently reproduced
85.8
GPT-4o (May 2024)
independently reproduced
84.67
GPT-4o (Nov 2024)
independently reproduced
84.13
GPT-4.1
independently reproduced
83.01
o1-preview
independently reproduced
81.65
GPT-5.3 Codex
independently reproduced
79.3
GPT-4o (Aug 2024)
independently reproduced
79.07
GPT-4.5
independently reproduced
78.63
GPT-4 Turbo (Apr 2024)
independently reproduced
75.07
GPT-5 Pro
independently reproduced
70.2
GPT-4.1 nano
independently reproduced
70
GPT-3.5 Turbo (Jun 2023)
unverified
63.2
GPT-3.5 Turbo (Jan 2024)
independently reproduced
56.4
gpt2
unverified
—
gpt-oss-20b
unverified
—
gpt-oss-120b
unverified
—
gpt2-large
unverified
—
447 source records · huggingface_models, epoch_benchmarks. Every score is evidence-linked.
Follow the record
Other labs
Qwen
Meta AI
DeepSeek
NVIDIA
Google DeepMind
EleutherAI
Zhipu AI
Hugging Face
Explore the record
Benchmark Integrity Index
Model Claim Ledger
Head-to-head comparisons