LongCat-Flash-Thinking

LongCat-Flash-Thinking ranks first on MATH-500.

Meituan Released 21 Sep 2025 Family LongCat-Flash The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Meituan Sep 2025

Open weights Flagship

No Epoch Capabilities Index score yet

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score	Rank	Source
MMLU-Pro	82.6%	#52 of 151	Model card
IFEval	86.9%	#57 of 126	Model card
Arena-Hard v2.0	69.9%	#12 of 24	Model card
MMLU-Redux	89.3%	#23 of 38	Model card
Extended NYT Connections	17.7%	#82 of 97	Official leaderboard

Benchmark	Score	Rank	Source
MATH-500	99.2%	#1 of 67	Model card
AIME 2024	93.3%	#6 of 87	Model card
ZebraLogic	95.5%	#3 of 23	Model card
AMO-Bench	43.6%	#4 of 21	Official leaderboard
AIME 2025	90.6%	#47 of 143	Model card
GPQA Diamond	81.5%	#100 of 284	Model card
Beyond AIME	69.5%	#5 of 14	Model card
HMMT 2025	83.7%	#32 of 49	Model card
miniF2F	81.6%	#4 of 5	Model card
ARC-AGI	50.3%	#18 of 19	Model card
General365	44.4%	#22 of 23	Official leaderboard
Sudoku-Bench	56%	#1 of 1	Model card

Benchmark	Score	Rank	Source
LiveCodeBench	79.4%	#15 of 73	Model card
LiveCodeBench v6	79.4%	#31 of 76	Model card
OJBench	40.7%	#6 of 12	Model card
SWE-bench Verified	59.4%	#110 of 150	Model card

Benchmark	Score	Rank	Source
Berkeley Function Calling Leaderboard (BFCL)	74.4%	#10 of 47	Model card
tau2-bench	74%	#45 of 69	Model card
VITA-Bench	29.5%	#10 of 12	Model card

Benchmark	Score		Rank	Source
LongCat safety evaluation (Meituan)	95.7%		#1 of 2	Model card

The benchmarks the lab chose to headline, with the scores it gave.

Every result in the longer documents the lab published with it. 1 of them isn't in the launch post.