Metr time horizons — Epoch AI

Metr time horizons — Epoch AI is run by Epoch AI. It has ranked 49 models, of which the catalogue holds 33, scoring from 0.101 to 0.789.

measured by Epoch AI · the board itself

Place	Model	Metric	Score
2ndof 49	Claude Opus 4.6	average_score	0.789
3rdof 49	Gemini 3.1 Pro Preview	average_score	0.77
4thof 49	GPT-5.2	average_score (high)	0.753
5thof 49	Claude Opus 4.5	average_score	0.75
6thof 49	GPT-5.3 Codex	average_score	0.745
7thof 49	GPT-5.4	average_score (xhigh)	0.743
9thof 49	gemini-3-pro-preview	average_score	0.71
10thof 49	GPT-5.1 Codex Max	average_score (max)	0.708
11thof 49	GPT-5	average_score (medium)	0.696
13thof 49	Claude Sonnet 4.5	average_score	0.674
14thof 49	Claude Opus 4.1	average_score	0.668
16thof 49	o3	average_score (medium)	0.654
17thof 49	Claude Opus 4	average_score	0.639
18thof 49	o4-mini	average_score (medium)	0.639
20thof 49	Claude Sonnet 4	average_score	0.62
23rdof 49	Claude 3.7 Sonnet	average_score	0.6
24thof 49	Kimi K2 Thinking	average_score	0.592
25thof 49	GPT OSS 120B	average_score	0.566
27thof 49	Gemini 2.5 Pro Preview 06-05	average_score	0.554
28thof 49	DeepSeek-R1 (0528)	average_score	0.538
29thof 49	DeepSeek-R1	average_score	0.519
30thof 49	o1	average_score (medium)	0.511
31stof 49	DeepSeek-V3 0324	average_score	0.496
32ndof 49	DeepSeek-V3	average_score	0.474
33rdof 49	Claude 3.5 Sonnet	average_score	0.452
35thof 49	GPT-4o (2024-05-13)	average_score	0.408
37thof 49	GPT-4 Turbo	average_score	0.367
39thof 49	Qwen2.5 72B Instruct	average_score	0.358
42ndof 49	Qwen2-72B-Instruct	average_score	0.299
43rdof 49	Claude 3 Opus	average_score	0.295
47thof 49	GPT-3.5 Turbo Instruct	average_score	0.215
48thof 49	davinci-002	average_score	0.162
49thof 49	gpt2-xl	average_score	0.101

Place

Model

Metric

Score

2ndof 49

average_score

0.789

3rdof 49

average_score

0.77