All makers

AI model maker

Meituan

4 models on Artificials, 1 with published test results and 1 with open weights you can download.

RSS feed
Best on the capability indexNot yetNone of its models is on the index yet
Best rankTop 17%LongCat Flash Chat 2602 Exp on Arena coding
Newest releaseLongCat 2.5 PreviewSep 25, 2026
Models41 tested, 1 with open weights

Where Meituan stands

Its best model’s rank among all models tested on each skill area’s main test, the same tests model pages use. A longer bar means a better rank.

  • OverallLongCat Flash Chat 2602 Exp on Arena text
    Top 23%
  • CodingLongCat Flash Chat 2602 Exp on Arena coding
    Top 17%
  • MathLongCat Flash Chat 2602 Exp on Arena math
    Top 23%
  • LanguageLongCat Flash Chat 2602 Exp on Arena creative writing
    Top 29%
  • ReasoningLongCat Flash Chat 2602 Exp on Arena hard prompts
    Top 22%

Every test

Its best result on each published test, and that result’s rank among every model tested. Scores from different tests are never combined.

TestBest modelScoreRank
Arena codingLongCat Flash Chat 2602 Exp1501#65 of 384
Arena hard promptsLongCat Flash Chat 2602 Exp1461#82 of 389
Arena textLongCat Flash Chat 2602 Exp1437#88 of 389
Arena mathLongCat Flash Chat 2602 Exp1435#85 of 372
Arena instruction followingLongCat Flash Chat 2602 Exp1416#100 of 389
Arena creative writingLongCat Flash Chat 2602 Exp1390#111 of 387

Every model

Newest first. The price is per 1M tokens, blended three input tokens to one output token: Meituan’s own listing, or else the middle price across hosts.

ModelReleasedContextPriceCapability index
LongCat 2.5 PreviewSep 25, 20261M tokens$0.525Not on the index
LongCat 2.5 Preview FreeSep 25, 20261M tokensNo paid listingNot on the index
LongCat-2.0Jun 30, 20261M tokens$1.30Not on the index
Meituan/Longcat-Flash-ChatNov 5, 2025131.1K tokensNo paid listingNot on the index

Models, prices, context windows and release dates from models.dev (MIT). Test results from Arena (CC BY 4.0). A model counts here when Meituan sells it, several hosts list it or a test covers it. Each figure is the source’s own; nothing is combined into a new score.