FrontierCode

Long, hard engineering tasks in large codebases. Published by Cognition.

Category	Coding
Metric	Accuracy
Source	Epoch AI (evaluator)
License	CC BY 4.0
Benchmark page	cognition.com
Status	Scored
Weight	1.3 within Coding
Fit	Discrimination 0.83, noise 0.18

Results

#	Model	Provider	Result	Date
1	Claude Opus 5.5 Medium	Anthropic	54.6%	2026-11-11
2	Claude Fable 5 Extra High	Anthropic	53.5%	2026-11-11
3	Claude Opus 5 Max	Anthropic	53.4%	2026-11-11
4	GPT-6 Astra Max	OpenAI	53.3%	2026-11-11
5	Claude Sonnet 5.5 Extra High	Anthropic	52.1%	2026-11-11
6	Claude Fable 5.1 Medium	Anthropic	50.9%	2026-11-11
7	GPT-6.1 Sol Medium	OpenAI	50.2%	2026-11-11
8	SWE-2 Max	Cognition	50.0%	2026-11-11
9	GPT-6 Sol Max	OpenAI	49.3%	2026-11-11
10	Grok 4.6 High	xAI	48.0%	2026-11-11
11	Grok 4.7 High	xAI	47.6%	2026-11-11
12	Gemini 4 Argon Medium	Google	47.6%	2026-11-11
13	GPT-5.6 Sol Max	OpenAI	47.5%	2026-11-11
14	Claude Opus 4.8 Max	Anthropic	46.5%	2026-11-11
15	Claude Haiku 5.5 Max	Anthropic	46.4%	2026-11-11
16	Kimi K3 Non-reasoning	Moonshot AI	44.2%	2026-11-11
17	DeepSeek V4.1 Flash High	DeepSeek	43.9%	2026-11-11
18	Gemini 3.7 Flash Medium	Google	43.6%	2026-11-11
19	GPT-5.5 Extra High	OpenAI	43.0%	2026-11-11
20	Claude Sonnet 5 Extra High	Anthropic	42.7%	2026-11-11
21	Grok 4.5 High	xAI	42.4%	2026-11-11
22	GPT-6 Luna Max	OpenAI	42.4%	2026-11-11
23	SWE-1.7 Non-reasoning	Cognition	42.0%	2026-11-11
24	GPT-5.6 Terra Max	OpenAI	41.3%	2026-11-11
25	Gemini 3.8 Flash Medium	Google	41.2%	2026-11-11
26	GLM-5.3 Max	Zhipu AI	40.1%	2026-11-11
27	GPT-5.6 Luna Max	OpenAI	39.8%	2026-11-11
28	Mistral Large 4 Preview High	Mistral	39.4%	2026-11-11
29	Claude Opus 4.7 Max	Anthropic	38.5%	2026-11-11
30	Gemini 3.6 Flash Medium	Google	34.4%	2026-11-11
31	GLM-5.3-Flash Max	Zhipu AI	31.8%	2026-11-11
32	Kimi K2.7 Code Non-reasoning	Moonshot AI	30.1%	2026-11-11
33	DeepSeek V4 Pro 0813 High	DeepSeek	28.6%	2026-11-11
34	GPT-5.4 mini Extra High	OpenAI	27.0%	2026-11-11
35	Claude Opus 4.6 High	Anthropic	26.6%	2026-11-11
36	Composer 2.5 Non-reasoning	Cursor	25.6%	2026-11-11
37	GLM-5.2 Non-reasoning	Zhipu AI	24.5%	2026-11-11
38	Claude Sonnet 4.6 Max	Anthropic	24.3%	2026-11-11
39	DeepSeek V4 Flash 0731 High	DeepSeek	18.8%	2026-11-11
40	DeepSeek V4 Pro Non-reasoning	DeepSeek	17.6%	2026-11-11
41	MiniMax-M3 Non-reasoning	MiniMax	14.7%	2026-11-11
42	Inkling	Thinking Machines	14.0%	2026-11-11
43	Nemotron 3 Ultra Non-reasoning	NVIDIA	13.6%	2026-11-11
44	Qwen3.7 Plus Non-reasoning	Alibaba	10.2%	2026-11-11
45	SWE-1.6 Non-reasoning	Cognition	9.4%	2026-11-11
46	Mistral Medium 3.5 Non-reasoning	Mistral	8.0%	2026-11-11