LiveBench Coding: JavaScriptBenchmark scores and sources

Published result; benchmark version and evaluation conditions remain in the id and result note.

pillar: coding · weight 1 within pillar · unit: % (higher is better) · official board ↗

Row	Model	Result	Normalized (0–100)	Confidence
1	DeepSeek V4.1 Flash DeepSeek	81.8%	81.8	69% confidence 69 percent, Medium
2	Qwen3.8 Max Alibaba / Qwen	77.3%	77.3	85% confidence 85 percent, High
3	Claude Opus 5 Anthropic	77.3%	77.3	100% confidence 100 percent, Full
4	Muse Spark 1.1 Meta	77.3%	77.3	61% confidence 61 percent, Medium
5	Muse Spark 1.3 Meta	77.3%	77.3	93% confidence 93 percent, High
6	Claude Opus 5.5 Anthropic	72.7%	72.7	100% confidence 100 percent, Full
7	Gemini 3.8 Flash Google	72.7%	72.7	100% confidence 100 percent, Full
8	Muse Spark 1.2 Meta	72.7%	72.7	53% confidence 53 percent, Medium
9	Grok 4.5 xAI	72.7%	72.7	100% confidence 100 percent, Full
10	Grok 4.6 xAI	72.7%	72.7	100% confidence 100 percent, Full
11	GLM-5.3 Z.ai	72.7%	72.7	93% confidence 93 percent, High
12	Qwen3.8 Flash Next Alibaba / Qwen	68.2%	68.2	21% confidence 21 percent, Low
13	Claude Fable 5 Anthropic	68.2%	68.2	100% confidence 100 percent, Full
14	Claude Fable 5.1 Anthropic	68.2%	68.2	100% confidence 100 percent, Full
15	Claude Haiku 5.5 Anthropic	68.2%	68.2	68% confidence 68 percent, Medium
16	Claude Opus 4.8 Anthropic	68.2%	68.2	100% confidence 100 percent, Full
17	Claude Sonnet 5 Anthropic	68.2%	68.2	93% confidence 93 percent, High
18	DeepSeek V4 Pro 0813 DeepSeek	68.2%	68.2	69% confidence 69 percent, Medium
19	Gemini 3.7 Flash Google	68.2%	68.2	100% confidence 100 percent, Full
20	Mistral Large 4 Mistral AI	68.2%	68.2	64% confidence 64 percent, Medium
21	Kimi K3 Moonshot AI	68.2%	68.2	100% confidence 100 percent, Full
22	GPT-5.2 Codex OpenAI	68.2%	68.2	43% confidence 43 percent, Low
23	GPT-5.4 OpenAI	68.2%	68.2	100% confidence 100 percent, Full
24	GPT-5.6 Terra OpenAI	68.2%	68.2	100% confidence 100 percent, Full
25	Inkling Thinking Machines Lab	68.2%	68.2	100% confidence 100 percent, Full
26	Claude Opus 4.6 Anthropic	63.6%	63.6	100% confidence 100 percent, Full
27	Claude Opus 4.7 Anthropic	63.6%	63.6	100% confidence 100 percent, Full
28	DeepSeek V4 Flash 0731 DeepSeek	63.6%	63.6	69% confidence 69 percent, Medium
29	DeepSeek V4 Flash Vision Exp DeepSeek	63.6%	63.6	21% confidence 21 percent, Low
30	Gemini 3.5 Flash Google	63.6%	63.6	100% confidence 100 percent, Full
31	Gemini 3.6 Flash Google	63.6%	63.6	100% confidence 100 percent, Full
32	MiniMax-M3 MiniMax	63.6%	63.6	100% confidence 100 percent, Full
33	Kimi K2.7 Code Moonshot AI	63.6%	63.6	48% confidence 48 percent, Low
34	GPT-5.4 nano OpenAI	63.6%	63.6	100% confidence 100 percent, Full
35	GPT-5.5 OpenAI	63.6%	63.6	100% confidence 100 percent, Full
36	GPT-5.6 Luna OpenAI	63.6%	63.6	100% confidence 100 percent, Full
37	GPT-5.6 Sol OpenAI	63.6%	63.6	100% confidence 100 percent, Full
38	GPT-6 Astra OpenAI	63.6%	63.6	100% confidence 100 percent, Full
39	GPT-6 Luna OpenAI	63.6%	63.6	100% confidence 100 percent, Full
40	GPT-6 Sol OpenAI	63.6%	63.6	100% confidence 100 percent, Full
41	GPT-6.1 Sol OpenAI	63.6%	63.6	100% confidence 100 percent, Full
42	Grok 4.7 xAI	63.6%	63.6	100% confidence 100 percent, Full
43	GLM-5.2 Z.ai	63.6%	63.6	100% confidence 100 percent, Full
44	GLM-5.3-Flash Z.ai	63.6%	63.6	100% confidence 100 percent, Full
45	Qwen3.6 Plus Alibaba / Qwen	59.1%	59.1	80% confidence 80 percent, High
46	Qwen3.7 Max Alibaba / Qwen	59.1%	59.1	53% confidence 53 percent, Medium
47	Qwen3.8 27B Alibaba / Qwen	59.1%	59.1	69% confidence 69 percent, Medium
48	Claude Opus 4.5 Anthropic	59.1%	59.1	100% confidence 100 percent, Full
49	Gemini 3.1 Pro Preview Google	59.1%	59.1	100% confidence 100 percent, Full
50	Gemini 3.5 Flash Lite Google	59.1%	59.1	100% confidence 100 percent, Full
51	Kimi K2.6 Moonshot AI	59.1%	59.1	85% confidence 85 percent, High
52	GPT-5.2 OpenAI	59.1%	59.1	100% confidence 100 percent, Full
53	Grok Build 0.1 xAI	59.1%	59.1	16% confidence 16 percent, Low
54	Qwen3.6 27B Alibaba / Qwen	54.5%	54.5	53% confidence 53 percent, Medium
55	Claude Sonnet 4.6 Anthropic	54.5%	54.5	100% confidence 100 percent, Full
56	DeepSeek V4 Flash DeepSeek	54.5%	54.5	53% confidence 53 percent, Medium
57	DeepSeek V4 Pro DeepSeek	54.5%	54.5	85% confidence 85 percent, High
58	Nemotron 3 Ultra 550B A55B NVIDIA	54.5%	54.5	100% confidence 100 percent, Full
59	GPT-5.4 mini OpenAI	50%	50.0	100% confidence 100 percent, Full
60	Claude Sonnet 5.5 Anthropic	36.4%	36.4	93% confidence 93 percent, High
61	Grok 4.3 xAI	27.3%	27.3	80% confidence 80 percent, High

Results are as published by the source behind each value (hover or tap the number). Benchmark scores are shown individually for every source/variant (a model may have multiple rows), and vary by version, harness and date; the normalized column uses the method’s fixed 0–100 scales and feeds the coding pillar of the SI Score.

LiveBench Coding: JavaScriptBenchmark scores and sources

What changed

Alerts on this device