Bbh — Epoch AI

Bbh — Epoch AI is run by Epoch AI. It has ranked 88 models, of which the catalogue holds 31, scoring from 0.288 to 0.875 on Average.

measured by Epoch AI · the board itself

Place	Model	Metric	Score
2ndof 88	DeepSeek-V3	Average	0.875
4thof 88	Llama 3.1 405B	Average	0.829
7thof 88	Phi-3-medium-128k-instruct	Average	0.814
10thof 88	Phi-3-small-8k-instruct	Average	0.791
11thof 88	DeepSeek-V2	Average	0.788
13thof 88	GPT-4 (0613)	Average	0.751
14thof 88	Phi-3-mini-4k-instruct	Average	0.717
15thof 88	Yi-34B-Chat	Average	0.717
16thof 88	StableBeluga2	Average	0.693
17thof 88	Llama-2-70b-hf	Average	0.649
18thof 88	GPT-3.5 Turbo (older v0613)	Average	0.616
19thof 88	phi-2	Average	0.594
21stof 88	Llama-2-70b-chat	Average	0.585
23rdof 88	Llama-2-13b-chat	Average	0.582
24thof 88	Mistral-7B-v0.1	Average	0.561
25thof 88	gemma-7b	Average	0.551
27thof 88	Qwen-14B-Chat	Average	0.55
28thof 88	Yi-34B	Average	0.543
29thof 88	Qwen-14B	Average	0.534
40thof 88	Baichuan2-13B-Chat	Average	0.472
42ndof 88	Llama-2-13b	Average	0.47
43rdof 88	Mistral 7B	Average	0.45
50thof 88	Baichuan-13B-Base	Average	0.43
51stof 88	Yi-6B	Average	0.428
62ndof 88	Llama-2-7b	Average	0.392
65thof 88	llama-13b	Average	0.379
67thof 88	falcon-40b	Average	0.371
72ndof 88	gemma-2b	Average	0.352
77thof 88	llama-7b	Average	0.335
81stof 88	Baichuan-7B	Average	0.325
85thof 88	falcon-7b	Average	0.288

Place

Model

Metric

Score

2ndof 88

Average

0.875

4thof 88

Average

0.829