Evaluation reports

We conduct evaluations of the autonomous capabilities of frontier AI models, with some in partnership with AI developers such as Anthropic and OpenAI. We do this both to understand the models' capabilities and to pilot third-party evaluator arrangements.

GPT-5.1-Codex-Max

19 November 2025 •
Partnership

GPT-5

7 August 2025 •
Partnership

DeepSeek and Qwen

27 June 2025 •
No company involvement

OpenAI o3 and o4-mini

16 April 2025 •
Partnership

Claude 3.7

4 April 2025 •
Partnership

DeepSeek-R1

5 March 2025 •
No company involvement

GPT-4.5

27 February 2025 •
Partnership

DeepSeek-V3

12 February 2025 •
No company involvement

Claude 3.5 Sonnet and o1

31 January 2025 •
Partnership

Claude 3.5 Sonnet (original)

30 October 2024 •
Partnership

o1-preview

12 September 2024 •
Partnership

GPT-4o

7 August 2024 •
Partnership

GPT-4 and Claude

17 March 2023 •
Partnership

METR does not accept compensation for this work.

Companies such as OpenAI and Anthropic have provided access and compute credits to support evaluation research. We also occasionally evaluate models independently after they are released, without involvement from the model's developers. Recent public reports resulting from this work are above, with additional discussion in the respective system cards.