51 models tracked · known for gpt2, gpt-oss-20b, gpt-oss-20b
Claims drawn from cited facts, not live model generation.
OpenAI is the lab that maintains a blog, described simply as 'The OpenAI blog'.
3 cited facts
OpenAI has shipped 51 models, establishing a substantial tracked footprint. Alongside that portfolio, OpenAI has made 1142 blog announcements, reflecting an actively communicating public presence. This mix of many shipped models and extensive blog activity signals a research-driven lab that engages its audience, not merely a quiet product shipper.
4 cited facts
566 benchmark scores across this lab’s models, tagged by evidence status.
| Model | Best score | Claim status |
|---|---|---|
| GPT-5.5 | 100 | independently reproduced |
| GPT-5.5 Pro | 100 | independently reproduced |
| GPT-5.6 Sol | 100 | independently reproduced |
| GPT-5.6 Terra | 99.72 | independently reproduced |
| GPT-5.6 Luna | 98.33 | independently reproduced |
| GPT-5 | 98.13 | independently reproduced |
| GPT-5 mini | 97.85 | independently reproduced |
| o4-mini | 97.83 | independently reproduced |
| GPT-5.4 | 97.78 | independently reproduced |
| o3 | 97.77 | independently reproduced |
| o3-pro | 97.2 | unverified |
| o3-mini | 96.49 | independently reproduced |
| GPT-5.2 | 96.11 | independently reproduced |
| GPT-5 nano | 95.24 | independently reproduced |
| o1 | 94.71 | independently reproduced |
| GPT-5.4 Pro | 94.5 | independently reproduced |
| GPT-4 (Mar 2023) | 93.73 | independently reproduced |
| GPT-4o mini | 91.3 | independently reproduced |
| GPT-5.2 Pro | 90.5 | independently reproduced |
| GPT-4 (Jun 2023) | 89.99 | independently reproduced |
| o1-mini | 89.18 | independently reproduced |
| gpt-oss-120b | 88.88 | independently reproduced |
| GPT-5.4 Mini | 88.88 | independently reproduced |
| GPT-5.1 | 88.6 | independently reproduced |
| GPT-5.4 Nano | 87.77 | independently reproduced |
| GPT-4.1 mini | 87.29 | independently reproduced |
| GPT-3.5 Turbo (Nov 2023) | 85.8 | independently reproduced |
| GPT-4o (May 2024) | 84.67 | independently reproduced |
| GPT-4o (Nov 2024) | 84.13 | independently reproduced |
| GPT-4.1 | 83.01 | independently reproduced |
| o1-preview | 81.65 | independently reproduced |
| GPT-5.3 Codex | 79.3 | independently reproduced |
| GPT-4o (Aug 2024) | 79.07 | independently reproduced |
| GPT-4.5 | 78.63 | independently reproduced |
| GPT-5.5 Instant | 76.68 | independently reproduced |
| GPT-4 Turbo (Apr 2024) | 75.07 | independently reproduced |
| GPT-4 Turbo (Nov 2023) | 72.8 | independently reproduced |
| GPT-5.1-Codex-Max | 71.58 | unverified |
| GPT-5 Pro | 70.2 | independently reproduced |
| GPT-4.1 nano | 70 | independently reproduced |
| GPT-3.5 Turbo (Jun 2023) | 63.2 | unverified |
| GPT-3.5 Turbo (Jan 2024) | 56.4 | independently reproduced |
| gpt-oss-20b | 53.84 | independently reproduced |
| gpt2 | — | unverified |
| gpt2-large | — | unverified |
| GPT-4o | — | unverified |
Where a model from OpenAI holds the top score, with the benchmark’s integrity grade.
Source facts, citations, and refresh stamp for this record.
Sources: curated_models, epoch_benchmarks, huggingface_models, lab_blog_rss