Skip to content

Benchmark

FrontierCode 1.1

Coding benchmarks compare models on programming tasks. Scores are percentages.

Models measured
41
Scores recorded
102
Sources
2
IQ weight
2.97

Score distribution

45.2%
Mean
47.5%
Median
12.5%
Std dev
27%
10th percentile
60.3%
90th percentile
9.3%
Min
9.3%65.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.97 for chat models, measured over 41 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model
1Claude Opus 5.5extended· FrontierCode data65.3%
2Claude Fable 5extended· FrontierCode data64.9%
3GPT-6 Astraextended· FrontierCode data64.5%
4Claude Sonnet 5.5extended· FrontierCode data64.4%
5Claude Opus 5extended· FrontierCode data63.6%
6Claude Fable 5.1extended· FrontierCode data63.6%
7Grok 4.6extended· FrontierCode data61.3%
8GPT-6 Solextended· FrontierCode data60.7%
9Gemini 4 Argonextended· FrontierCode data60.6%
10GPT-5.6 Solextended· FrontierCode data60.5%
11GPT-6.1 Solextended· FrontierCode data60.4%
12Claude Opus 4.8extended· FrontierCode data59.6%
13Grok 4.7extended· FrontierCode data59.4%
14Claude Haiku 5.5extended· FrontierCode data58.4%
15Kimi K3extended· FrontierCode data58.2%
16GPT 5.5extended· FrontierCode data56.7%
17DeepSeek V4.1 Flashextended· FrontierCode data56.6%
18Grok 4.5extended· FrontierCode data56.6%
19Gemini 3.7 Flashextended· FrontierCode data56.3%
20Claude Sonnet 5extended· FrontierCode data56.2%
21GPT-6 Lunaextended· FrontierCode data56.1%
22GPT-5.6 Terraextended· FrontierCode data55.8%
23GPT-5.6 Lunaextended· FrontierCode data55.1%
24Claude Opus 5.5main· FrontierCode data54.6%
25Claude Opus 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim54.4%
26Claude Opus 4.7extended· FrontierCode data53.9%
27Claude Fable 5· LLM Stats data53.5%
28Claude Fable 5main· FrontierCode data53.5%
29Gemini 3.8 Flashextended· FrontierCode data53.5%
30Claude Opus 5· LLM Stats data53.4%
31Claude Opus 5main· FrontierCode data53.4%
32GPT-6 Astra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim53.3%
33GPT-6 Astramain· FrontierCode data53.3%
34GLM 5.3extended· FrontierCode data52.6%
35Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim52.1%
36Claude Sonnet 5.5main· FrontierCode data52.1%
37Claude Fable 5.1main· FrontierCode data50.9%
38GPT-6.1 Solmain· FrontierCode data50.2%
39GPT-6.1 Sol· LLM Stats data50.2%
40GPT-6 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim49.3%
41GPT-6 Solmain· FrontierCode data49.3%
42Gemini 3.6 Flashextended· FrontierCode data48%
43Grok 4.6main· FrontierCode data48%
44Grok 4.7main· FrontierCode data47.6%
45Gemini 4 Argonmain· FrontierCode data47.6%
46GPT-5.6 Sol· LLM Stats data47.5%
47GPT-5.6 Solmain· FrontierCode data47.5%
48Claude Opus 4.8· LLM Stats data46.5%
49Claude Opus 4.8main· FrontierCode data46.5%
50Claude Haiku 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim46.4%

Showing 1–50 of 102 results.

Tokens the model generated per task, as measured by the source. Comparable within this benchmark only.

Data credits: Data by LLM Stats