LiveBench Math

Name: LiveBench Math
Creator: aib vote
License: https://www.gnu.org/licenses/gpl-3.0.html

About This Benchmark

LiveBench Math category score (0–100). Evaluates competition-level math problems, proofs, and numerical computation.

Rank	Model
#1	OpenAI GPT-5.5	96.3
#2	OpenAI GPT-5.4	94.2
#3	Google Gemini 3.1 Pro	91.0
#4	Anthropic Claude Opus 4.6	89.3
#5	Grok Grok 4.20 (Reasoning)	87.1
#6	OpenAI GPT-5	86.2
#7	Z.ai GLM-5.1	84.9
#8	Moonshot AI Kimi K2.5	84.9
#9	Moonshot AI Kimi K2.6	84.3
#10	Grok Grok 4.1 Fast (Reasoning)	83.7
#11	Alibaba Qwen3.6 Plus	83.7
#12	Z.ai GLM-5	83.5
#13	MiniMax MiniMax M2.7	80.5
#14	Anthropic Claude Sonnet 4.6	80.4
#15	MiniMax MiniMax M2.5	77.4
#16	Xiaomi MiMo-V2-Pro	77.0
#17	Anthropic Claude Opus 4.7	76.3
#18	OpenAI GPT-5 Mini	74.4
#19	Google Gemma 4 31B	73.9
#20	Google Gemini 3.1 Flash Lite	73.6
#21	Anthropic Claude Opus 4.1	73.2
#22	Anthropic Claude Sonnet 4	70.5
#23	OpenAI GPT OSS 120B	68.9
#24	Google Gemini 2.5 Flash	68.8
#25	Google Gemini 2.5 Pro	68.3
#26	Google Gemini 3 Flash	68.1
#27	OpenAI GPT-5 Nano	64.7
#28	Anthropic Claude Opus 4.5	64.1
#29	DeepSeek DeepSeek V3.2	64.0
#30	Anthropic Claude Sonnet 4.5	62.6
#31	Google Gemini 2.5 Flash Lite	61.0
#32	Anthropic Claude Haiku 4.5	58.0
#33	Grok Grok 4.20	45.5
#34	Arcee AI Trinity Large Thinking	44.9
#35	Grok Grok 4.1 Fast	38.9
#36	OpenAI GPT-5.4 Mini	37.0
#37	NVIDIA Nemotron 3 Super	36.4
#38	OpenAI GPT-5.4 Nano	36.0