Coding

FrontierCode 2.0 leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap
License: https://creativecommons.org/licenses/by-nc/4.0/

As of 2026-10-29, the highest measured score on FrontierCode 2.0 is 45.5% by Claude Opus 5.5. 91 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 5.5	45.5%	measured
2	GPT-6 Astra	42.8%	measured
3	Claude Sonnet 5.5	41.0%	measured
4	GPT-6.1 Sol	39.0%	measured
5	Gemini 4 Carbon	38.6%	estimated ± 4.9 pp, high confidence
6	GPT-6 Sol	34.3%	measured
7	Muse Wonder Full	33.9%	estimated ± 7.0 pp, low confidence
8	Claude Mythos 5	33.0%	estimated ± 7.1 pp, low confidence
9	Claude Fable 5.1	32.4%	measured
10	Step 5 Preview	31.5%	estimated ± 10.1 pp, low confidence
11	Kimi K3.1 Pro	30.4%	estimated ± 7.0 pp, low confidence
12	MiniMax M3.1 Flash	29.0%	estimated ± 7.6 pp, medium confidence
13	Grok 4.7	28.6%	measured
14	Claude Haiku 5.5	27.7%	measured
15	GPT-5.6 Sol	27.6%	measured
16	Sakana Fugu-Ultra	27.2%	estimated ± 7.1 pp, low confidence
17	Claude Fable 5	27.0%	estimated ± 4.9 pp, high confidence
18	Claude Opus 5	26.7%	measured
19	Gemini 4 Argon	26.3%	measured
20	Muse Spark 1.3	25.1%	estimated ± 2.6 pp, low confidence
21	Grok 4.6	25.0%	estimated ± 2.6 pp, low confidence
22	GPT-5.6 Terra	24.9%	estimated ± 2.6 pp, low confidence
23	Muse Wonder	24.3%	measured
24	Grok 4.5	24.2%	estimated ± 5.2 pp, low confidence
25	Gemini 3.8 Flash	24.1%	estimated ± 2.6 pp, low confidence
26	Ling 3.1 Flash	24.0%	estimated ± 7.1 pp, low confidence
27	SWE-2	23.8%	measured
28	GLM-5.3	23.7%	estimated ± 4.9 pp, high confidence
29	Gemini 3.7 Flash	23.5%	estimated ± 4.9 pp, medium confidence
30	GLM-5.3-Flash	23.5%	estimated ± 4.9 pp, medium confidence
31	Qwen3.8 Max	23.5%	estimated ± 4.9 pp, medium confidence
32	Muse Spark 1.2	23.5%	estimated ± 4.9 pp, medium confidence
33	Inkling	23.5%	estimated ± 4.9 pp, medium confidence
34	GPT-5.6 Luna	22.8%	estimated ± 2.6 pp, low confidence
35	MiMo-V2.6-Pro	22.7%	estimated ± 7.0 pp, low confidence
36	GPT-5.5	22.6%	measured
37	Beam 1.0 501B-A23B	22.6%	estimated ± 7.1 pp, low confidence
38	Beam 1.0 501B-A23B FP8	22.5%	estimated ± 7.1 pp, low confidence
39	Hy4 preview	22.4%	estimated ± 7.1 pp, low confidence
40	Beam 1.0 501B-A23B NVFP4	22.4%	estimated ± 7.1 pp, low confidence
41	Claude Sonnet 5	22.3%	estimated ± 2.6 pp, low confidence
42	Ornith-1.5-397B	22.1%	estimated ± 7.1 pp, low confidence
43	GPT-6 Luna	21.9%	measured
44	Claude Opus 4.7 (Adaptive)	21.8%	estimated ± 7.1 pp, low confidence
45	Composer 2.5	21.4%	estimated ± 2.6 pp, low confidence
46	Qwen3.8-Omni-Flash	21.4%	estimated ± 7.1 pp, low confidence
47	Qwen3.8-Flash-Next	21.1%	estimated ± 7.1 pp, low confidence
48	Ornith-1.0-397B	21.0%	estimated ± 7.1 pp, low confidence
49	Muse Spark 1.1	20.7%	estimated ± 7.1 pp, low confidence
50	Sage-A1.5	20.6%	estimated ± 7.1 pp, low confidence
51	dots3-note Preview	20.5%	estimated ± 7.1 pp, low confidence
52	Qwen3.7 Max	20.4%	estimated ± 7.1 pp, low confidence
53	Atria Dawn Preview	20.1%	estimated ± 7.1 pp, low confidence
54	Ornith-1.5-35B-A3B	20.1%	estimated ± 7.1 pp, low confidence
55	Laguna S 2.1	20.0%	estimated ± 7.1 pp, low confidence
56	MiniMax M3	19.9%	estimated ± 7.1 pp, low confidence
57	Sakana Fugu	19.9%	estimated ± 7.1 pp, low confidence
58	Kimi K2.6	19.8%	estimated ± 7.1 pp, low confidence
59	GLM-5.1	19.7%	estimated ± 7.1 pp, low confidence
60	Mistral Large 4	19.7%	measured
61	GPT-5.4	19.6%	estimated ± 7.1 pp, low confidence
62	Qwen3.7 Plus	19.5%	estimated ± 7.1 pp, low confidence
63	Qwen 3.6 Max (preview)	19.5%	estimated ± 7.1 pp, low confidence
64	MiMo-V2.5-Pro	19.4%	estimated ± 7.1 pp, low confidence
65	Claude Opus 4.5	19.4%	estimated ± 7.1 pp, low confidence
66	GPT-5.3 Codex	19.3%	estimated ± 7.1 pp, low confidence
67	DeepSeek V4.1 Flash 1017	19.3%	measured
68	Ling 3.0 Flash	19.3%	estimated ± 7.1 pp, low confidence
69	Qwen3.6 Plus	19.3%	estimated ± 7.1 pp, low confidence
70	Step 3.7 Flash	19.2%	estimated ± 7.1 pp, low confidence
71	MiniMax M2.7	19.2%	estimated ± 7.1 pp, low confidence
72	MiMo-V2.5	19.2%	estimated ± 7.1 pp, low confidence
73	Inkling-Small	19.1%	estimated ± 7.1 pp, low confidence
74	GPT-5.2	19.1%	estimated ± 7.1 pp, low confidence
75	DeepSeek V4 Pro 0813	19.0%	estimated ± 7.1 pp, low confidence
76	GLM-5	18.9%	estimated ± 7.1 pp, low confidence
77	Spark-X2.6-4B	18.8%	estimated ± 7.1 pp, low confidence
78	Gemini 3.5 Flash-Lite	18.8%	estimated ± 7.1 pp, low confidence
79	Qwen3.6-27B	18.6%	estimated ± 7.1 pp, low confidence
80	Claude Opus 4.6	18.6%	estimated ± 7.1 pp, low confidence
81	MAI-Thinking-1	18.5%	estimated ± 7.1 pp, low confidence
82	DeepSeek V4 Flash 0731	18.5%	estimated ± 7.1 pp, low confidence
83	Muse Spark	18.5%	estimated ± 7.1 pp, low confidence
84	Grok 4.20	18.4%	estimated ± 7.1 pp, low confidence
85	Nemotron 3.5 Super VL 120B A12B	18.4%	estimated ± 7.1 pp, low confidence
86	Nemotron 3.5 Super VL 120B A12B NVFP4	18.3%	estimated ± 7.1 pp, low confidence
87	Muse Glimmer 30B	18.3%	estimated ± 7.1 pp, low confidence
88	Qwen3.8-27B	18.3%	estimated ± 7.0 pp, low confidence
89	Qwen3.5 397B	18.2%	estimated ± 7.1 pp, low confidence
90	Kimi K2.5	18.2%	estimated ± 7.1 pp, low confidence
91	Ornith-1.0-35B	18.2%	estimated ± 7.1 pp, low confidence
92	Qwen3.6-35B-A3B	18.0%	estimated ± 7.1 pp, low confidence
93	Claude Opus 4.7	18.0%	estimated ± 7.6 pp, medium confidence
94	Laguna M.1	18.0%	estimated ± 7.1 pp, low confidence
95	MiMo-V2.6-Flash	17.9%	estimated ± 10.1 pp, low confidence
96	Laguna XS 2.1	17.8%	estimated ± 7.1 pp, low confidence
97	Ornith-1.5-9B	17.8%	estimated ± 7.1 pp, low confidence
98	Kimi K3	17.7%	measured
99	Laguna XS.2	17.7%	estimated ± 7.1 pp, low confidence
100	Ornith-1.0-9B	17.4%	estimated ± 7.1 pp, low confidence
101	LongCat-Flash-Lite-Sparse	17.3%	estimated ± 7.1 pp, low confidence
102	Granite 4.2 30B	17.1%	estimated ± 7.1 pp, low confidence
103	LLaDA2.2-flash	17.1%	estimated ± 7.1 pp, low confidence
104	Mellum2.1-12B-A2.5B-Thinking	17.0%	estimated ± 7.1 pp, low confidence
105	SWE-1.7	16.2%	measured
106	GLM-5.2	16.1%	estimated ± 5.2 pp, low confidence
107	Gemini 3.6 Flash	15.5%	estimated ± 5.2 pp, low confidence
108	Claude Opus 4.8	14.7%	measured
109	Kimi K2.7 Code	14.3%	estimated ± 5.2 pp, low confidence
110	Gemini 3.5 Flash	14.0%	estimated ± 5.2 pp, low confidence
111	Gemini 3.1 Pro	12.7%	estimated ± 7.6 pp, low confidence

FrontierCode 2.0 leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General