CodePanel

JavaScript code review: which models catch the bugs

Of the 17 models ranked on JavaScript code, 4 models tied for the most known bugs: 100%. 7 of the other 13 models have 95% intervals that overlap theirs. Gemma 4 26B A4B and Gemini 3.5 Flash-Lite tied for the fewest, 0%.

Updated 2026-11-04Full leaderboardMethodology

Results

Known defects caught, with 95% intervals. The tick is the model's recall over every bug.

JavaScript code review: which models catch the bugs
ModelCaught hereEvery bug
Claude Opus 5.5100%90%
GPT-6.1 Sol100%89%
Claude Sonnet 5.5100%77%
GPT-5.6 Sol100%72%
Gemini 4 Argon67%[0%–100%]82%
Gemini 3.8 Flash67%[0%–100%]68%
GPT-6 Luna67%[0%–100%]67%
GLM-5.3 Flash (Workers AI)67%[33%–100%]61%
DeepSeek V4 Flash (Workers AI)56%[0%–100%]49%
Claude Haiku 5.533%[0%–100%]64%
Gemini 3.1 Pro33%[0%–100%]67%
GLM-5.3 (Workers AI)33%[0%–67%]61%
DeepSeek V4 Pro (Workers AI)17%[0%–50%]54%
GPT-OSS 120B (Workers AI)11%[0%–33%]40%
Kimi K2.7 Code (Workers AI)8%[0%–25%]38%
Gemma 4 26B A4B (Workers AI)0%26%
Gemini 3.5 Flash-Lite0%11%

Against the rest of the benchmark

On average these models caught 50% of known bugs in JavaScript code, less often than the 60% the same 17 caught across the whole benchmark: the second-hardest of the 12 languages.

GPT-5.6 Sol gains the most here: 100%, against 72% over every bug.

DeepSeek V4 Pro (Workers AI) drops the most: 17%, against 54% over every bug.

A test PR that split the models

The change “Add `weight` option” in sindresorhus/p-throttle, a change of 100 to 300 lines that landed in November 2025, introduced a logic bug that a later upstream commit fixed.

Claude Opus 5.5, GPT-6.1 Sol, Gemini 4 Argon, Claude Sonnet 5.5, GPT-5.6 Sol and Gemini 3.8 Flash caught it; GLM-5.3 Flash caught it in part; GPT-6 Luna, Claude Haiku 5.5, Gemini 3.1 Pro, GLM-5.3, DeepSeek V4 Pro (Workers AI), DeepSeek V4 Flash (Workers AI), GPT-OSS 120B (Workers AI), Kimi K2.7 Code (Workers AI), Gemma 4 26B A4B and Gemini 3.5 Flash-Lite missed it.