LiveBench

Contamination-resistant general capability, refreshed monthly by the LiveBench project.

#	Model	Maker	Score	Measured	Link
1	Claude Fable 5.1	Anthropic	88.0 / 100	Oct 14, 2026	Source
2	Claude Opus 5.5	Anthropic	86.5 / 100	Oct 14, 2026	Source
3	GPT-6 Astra	OpenAI	86.1 / 100	Oct 14, 2026	Source
4	GPT-6.1 Sol	OpenAI	85.4 / 100	Oct 14, 2026	Source
5	Gemini 4 Argon	Google DeepMind	84.1 / 100	Oct 14, 2026	Source
6	Claude Sonnet 5.5	Anthropic	82.5 / 100	Oct 14, 2026	Source
7	Claude Opus 5	Anthropic	82.1 / 100	Oct 14, 2026	Source
8	Muse Wonder	Meta	78.1 / 100	Oct 15, 2026	Source
9	DeepSeek V4.1 Flash	DeepSeek	77.3 / 100	Oct 14, 2026	Source
10	GPT-5.6 Sol	OpenAI	76.3 / 100	Oct 14, 2026	Source
11	GPT-5.6 Terra	OpenAI	75.0 / 100	Oct 14, 2026	Source
12	Muse Spark 1.3	Meta	74.9 / 100	Oct 14, 2026	Source
13	Claude Sonnet 5	Anthropic	74.6 / 100	Oct 14, 2026	Source
14	Claude Haiku 5.5	Anthropic	74.5 / 100	Oct 14, 2026	Source
15	Grok 4.7	xAI (SpaceXAI)	74.1 / 100	Oct 14, 2026	Source
16	GPT-6 Luna	OpenAI	69.4 / 100	Oct 14, 2026	Source
17	Gemini 3.8 Flash	Google DeepMind	67.3 / 100	Oct 14, 2026	Source
18	GPT-5.6 Luna	OpenAI	66.1 / 100	Oct 14, 2026	Source
19	Qwen3.8-Max	Alibaba Qwen	64.2 / 100	Oct 14, 2026	Source
20	Kimi K3	Moonshot AI	63.4 / 100	Oct 14, 2026	Source
21	Gemini 3.1 Pro Preview	Google DeepMind	54.9 / 100	Oct 14, 2026	Source
22	Mistral Large 4	Mistral AI	47.7 / 100	Oct 15, 2026	Source
23	Claude Haiku 4.5	Anthropic	46.0 / 100	Oct 14, 2026	Source

Scores are the benchmark owners’ published results, recorded with the date we verified them. This benchmark feeds the League Score — see methodology.