benchgap
Coding

FrontierCode 1.1 Extended leaderboard

As of 2026-11-21, the highest measured score on FrontierCode 1.1 Extended is 64.5% by GPT-6 Astra. 90 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1GPT-6 Astra64.5%measured
2Claude Fable 5.264.0%estimated ± 3.0 pp, low confidence
3Claude Opus 563.6%measured
4Claude Opus 5.563.6%measured
5Gemini 4 Carbon62.5%estimated ± 3.0 pp, low confidence
6Claude Fable 5.162.5%estimated ± 1.6 pp, low confidence
7Claude Fable 562.4%estimated ± 1.6 pp, medium confidence
8Gemini 3.8 Flash62.3%estimated ± 1.6 pp, medium confidence
9Gemini 4 Hydrogen62.1%estimated ± 3.0 pp, medium confidence
10Gemini 4 Argon62.0%estimated ± 3.0 pp, medium confidence
11Muse Wonder61.7%estimated ± 3.0 pp, medium confidence
12Claude Mythos 561.5%estimated ± 3.0 pp, medium confidence
13Claude Haiku 5.561.4%estimated ± 3.0 pp, medium confidence
14DeepSeek V4.1 Flash 101761.3%estimated ± 3.0 pp, medium confidence
15Grok 4.661.3%measured
16DeepSeek V4 Pro61.1%estimated ± 2.4 pp, medium confidence
17Grok 4.761.0%estimated ± 3.0 pp, medium confidence
18GPT-5.6 Sol60.6%measured
19Muse Spark 1.260.3%estimated ± 3.4 pp, low confidence
20Gemini 3.7 Flash60.1%estimated ± 3.4 pp, low confidence
21Sakana Fugu-Ultra59.9%estimated ± 3.0 pp, medium confidence
22Muse Spark 1.359.5%estimated ± 3.0 pp, medium confidence
23Grok 4.559.4%estimated ± 1.6 pp, medium confidence
24GLM-5.359.2%estimated ± 2.4 pp, medium confidence
25Claude Sonnet 5.559.1%measured
26Ling 3.1 Flash58.6%estimated ± 3.0 pp, medium confidence
27Qwen3.8 Max58.3%estimated ± 3.0 pp, medium confidence
28Beam 1.0 501B-A23B57.9%estimated ± 3.0 pp, medium confidence
29Beam 1.0 501B-A23B FP857.8%estimated ± 3.0 pp, medium confidence
30Hy4 preview57.8%estimated ± 3.0 pp, medium confidence
31Beam 1.0 501B-A23B NVFP457.7%estimated ± 3.0 pp, medium confidence
32Qwen3.8 Max Preview57.7%estimated ± 3.4 pp, low confidence
33Ornith-1.5-397B57.6%estimated ± 3.0 pp, medium confidence
34Claude Opus 4.757.4%estimated ± 3.0 pp, medium confidence
35Qwen3.8-Omni-Flash57.1%estimated ± 3.0 pp, medium confidence
36Qwen3.8-Flash-Next56.8%estimated ± 3.0 pp, low confidence
37Ornith-1.0-397B56.7%estimated ± 3.0 pp, low confidence
38Qwen3.8-27B56.5%estimated ± 3.0 pp, low confidence
39Muse Spark 1.156.5%estimated ± 3.0 pp, low confidence
40Ripple-65B Preview56.5%estimated ± 3.0 pp, low confidence
41Sage-A1.556.4%estimated ± 3.0 pp, low confidence
42dots3-note Preview56.3%estimated ± 3.0 pp, low confidence
43Qwen3.7 Max56.2%estimated ± 3.0 pp, low confidence
44Gemini 3.1 Pro56.0%estimated ± 3.4 pp, low confidence
45Atria Dawn Preview55.9%estimated ± 3.0 pp, low confidence
46Ornith-1.5-35B-A3B55.9%estimated ± 3.0 pp, low confidence
47Laguna S 2.155.8%estimated ± 3.0 pp, low confidence
48GPT-5.6 Terra55.8%measured
49MiniMax M355.7%estimated ± 3.0 pp, low confidence
50Sakana Fugu55.7%estimated ± 3.0 pp, low confidence
51Kimi K2.655.5%estimated ± 3.0 pp, low confidence
52GLM-5.155.5%estimated ± 3.0 pp, low confidence
53GPT-5.455.2%estimated ± 3.0 pp, low confidence
54Qwen3.7 Plus55.2%estimated ± 3.0 pp, low confidence
55Claude Opus 4.855.2%estimated ± 1.6 pp, medium confidence
56Claude Sonnet 555.1%estimated ± 1.6 pp, medium confidence
57Kimi K355.1%estimated ± 1.6 pp, low confidence
58GPT-5.555.1%estimated ± 1.6 pp, low confidence
59Composer 2.555.1%estimated ± 1.6 pp, low confidence
60GLM-5.255.1%estimated ± 1.6 pp, low confidence
61Gemini 3.6 Flash55.1%estimated ± 1.6 pp, low confidence
62GPT-5.6 Luna55.1%measured
63Qwen 3.6 Max (preview)55.1%estimated ± 3.0 pp, low confidence
64MiMo-V2.5-Pro55.0%estimated ± 3.0 pp, low confidence
65Claude Opus 4.555.0%estimated ± 3.0 pp, low confidence
66GPT-5.3 Codex54.9%estimated ± 3.0 pp, low confidence
67Ling 3.0 Flash54.8%estimated ± 3.0 pp, low confidence
68Qwen3.6 Plus54.8%estimated ± 3.0 pp, low confidence
69Step 3.7 Flash54.7%estimated ± 3.0 pp, low confidence
70MiniMax M2.754.7%estimated ± 3.0 pp, low confidence
71MiMo-V2.554.7%estimated ± 3.0 pp, low confidence
72Inkling-Small54.6%estimated ± 3.0 pp, low confidence
73GPT-5.254.5%estimated ± 3.0 pp, low confidence
74Gemini 3.5 Flash54.3%estimated ± 3.0 pp, low confidence
75GLM-554.3%estimated ± 3.0 pp, low confidence
76Inkling54.0%estimated ± 3.0 pp, low confidence
77Spark-X2.6-4B54.0%estimated ± 3.0 pp, low confidence
78Gemini 3.5 Flash-Lite54.0%estimated ± 3.0 pp, low confidence
79Qwen3.6-27B53.7%estimated ± 3.0 pp, low confidence
80Claude Opus 4.653.7%estimated ± 3.0 pp, low confidence
81MAI-Thinking-153.4%estimated ± 3.0 pp, low confidence
82DeepSeek V4 Flash 073153.3%estimated ± 3.0 pp, low confidence
83Muse Spark53.3%estimated ± 3.0 pp, low confidence
84Grok 4.2053.0%estimated ± 3.0 pp, low confidence
85Nemotron 3.5 Super VL 120B A12B53.0%estimated ± 3.0 pp, low confidence
86Nemotron 3.5 Super VL 120B A12B NVFP452.8%estimated ± 3.0 pp, low confidence
87Muse Glimmer 30B52.8%estimated ± 3.0 pp, low confidence
88GLM-5.3-Flash52.7%estimated ± 2.4 pp, low confidence
89Qwen3.5 397B52.7%estimated ± 3.0 pp, low confidence
90Kimi K2.552.6%estimated ± 3.0 pp, low confidence
91Ornith-1.0-35B52.5%estimated ± 3.0 pp, low confidence
92Qwen3.6-35B-A3B52.1%estimated ± 3.0 pp, low confidence
93Laguna M.152.0%estimated ± 3.0 pp, low confidence
94Laguna XS 2.151.3%estimated ± 3.0 pp, low confidence
95Ornith-1.5-9B51.2%estimated ± 3.0 pp, low confidence
96Laguna XS.250.7%estimated ± 3.0 pp, low confidence
97Ornith-1.0-9B49.1%estimated ± 3.0 pp, low confidence
98LongCat-Flash-Lite-Sparse47.9%estimated ± 3.0 pp, low confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General