Arc ai2 — Epoch AI

Arc ai2 — Epoch AI is run by Epoch AI. It has ranked 134 models, of which the catalogue holds 38, scoring from 0.25 to 0.953 on Challenge score.

measured by Epoch AI · the board itself

Place	Model	Metric	Score
1stof 134	Llama 3.1 405B	Challenge score	0.953
2ndof 134	DeepSeek-V3	Challenge score	0.953
4thof 134	DeepSeek-V2	Challenge score	0.922
5thof 134	Phi-3-medium-128k-instruct	Challenge score	0.916
6thof 134	Phi-3-small-8k-instruct	Challenge score	0.907
7thof 134	GPT-3.5 Turbo (1106)	Challenge score	0.874
10thof 134	StableBeluga2	Challenge score	0.861
14thof 134	Phi-3-mini-4k-instruct	Challenge score	0.849
15thof 134	Qwen-14B	Challenge score	0.844
17thof 134	Llama 3 8B Instruct	Challenge score	0.828
19thof 134	Mistral-7B-v0.1	Challenge score	0.786
20thof 134	Llama-2-70b-hf	Challenge score	0.783
21stof 134	gemma-7b	Challenge score	0.783
22ndof 134	phi-2	Challenge score	0.759
23rdof 134	Mistral 7B	Challenge score	0.753
24thof 134	Qwen 2.5 Coder 32B	Challenge score	0.705
29thof 134	falcon-180B	Challenge score	0.678
36thof 134	falcon-40b	Challenge score	0.619
38thof 134	Qwen2.5-Coder-7B	Challenge score	0.609
39thof 134	Llama-2-13b	Challenge score	0.603
54thof 134	Yi-9B	Challenge score	0.556
73rdof 134	llama-13b	Challenge score	0.527
82ndof 134	Yi-6B	Challenge score	0.503
88thof 134	falcon-7b	Challenge score	0.479
89thof 134	llama-7b	Challenge score	0.476
92ndof 134	starcoder2-15b	Challenge score	0.472
93rdof 134	Llama-2-7b	Challenge score	0.459
100thof 134	phi-1_5	Challenge score	0.444
111thof 134	gemma-2b	Challenge score	0.421
113thof 134	xgen-7b-8k-base	Challenge score	0.412
114thof 134	gpt-neox-20b	Challenge score	0.411
118thof 134	starcoder2-7b	Challenge score	0.387
124thof 134	gpt-j-6b	Challenge score	0.363
125thof 134	opt-1.3b	Challenge score	0.358
127thof 134	Qwen2.5-Coder-0.5B	Challenge score	0.344
128thof 134	starcoder2-3b	Challenge score	0.342
131stof 134	stablelm-tuned-alpha-7b	Challenge score	0.27
133rdof 134	gpt2-xl	Challenge score	0.25

Place

Model

Metric

Score

1stof 134

Llama 3.1 405B

Challenge score

0.953

2ndof 134

DeepSeek-V3

Challenge score

0.953