LongCatReleased Nov 2, 2026
NewProprietary
LongCat 2.5
Strongest in Computer Use (#23 of 180). Six of seven dimensions are estimated, because no results are published yet.
Composite IQ ?
105
Effective cost
Not priced
No standard API rate
Speed
Not measured
No speed data yet
Context
Unverified
Not in publisher docs
Evidence
1of 7 dims
1 of 39 benchmarks
IQ profile
Seven dimensions, averaged equally into Composite IQ
MeasuredEstimated95% range
#1 in dimensionTop-10 median
IQ dimensions
Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.
DimensionIQPosition · 70–160RankEvidencevs Top 10
LongCat 2.595% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human
Benchmark results
1 of 39 Composite IQ benchmarks have a published result for LongCat 2.5. Benchmarks without a result stay listed so gaps are visible.
Abstract Reasoning IQ 85 · estimatedNo published results0/5
No published results yet for Abstract Reasoning. The 85 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
ARC-AGI-3ARC-AGI-2ARC-AGI-1Chess PuzzlesMystery Game Puzzles
Mathematical Reasoning IQ 94 · estimatedNo published results0/5
No published results yet for Mathematical Reasoning. The 94 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1MathArenaAIME
Academic Reasoning IQ 118 · estimatedNo published results0/6
No published results yet for Academic Reasoning. The 118 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
Humanity's Last ExamGPQA DiamondCritPtSciCodeMMLU-ProMMMU-Pro
Programmatic Reasoning IQ 99 · estimatedNo published results0/6
No published results yet for Programmatic Reasoning. The 99 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
LiveCodeBenchIOI V2Terminal-Bench 4.0ProgramBenchSWE-rebenchMirrorCode
Computer Use IQ 127 · measuredBest: #2 of 45 on BrowseComp1/5
BenchmarkScorePercentileRankLeader · source
BrowseComp
Full source leaderboard reconciliation · Nov 13
91.9%
#2 / 45
Leader Kimi K3.1 Pro · 92.1%
Full source leaderboard reconciliation · Nov 13
Full source leaderboard reconciliation · Nov 13
Not yet reported
OSWorld-VerifiedMCP AtlasArena.ai Agent ArenaAgents' Last Exam
Reliability IQ 107 · estimatedNo published results0/6
No published results yet for Reliability. The 107 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
SimpleQA VerifiedaaOmniscienceBullshitBench v2IFBenchAA Long Context Reasoning v1.1FACTS Grounding
Knowledge Work IQ 108 · estimatedNo published results0/6
No published results yet for Knowledge Work. The 108 shown is estimated from LongCat 2.5's other dimensions and will be replaced as results come in. How estimates work
AA-Briefcase v1.1GDPval-AA v2.1AutomationBench-AAGDP.pdfAA-AnalystAgentEnterpriseOps-Gym-AA
Release & availability
Facts from the publisher's announcement and documentation.
- Released
- Developer
- LongCat
- Access
- Meituan launched LongCat-2.5 on November 2, 2026. It is available through LongCat Chat and the LongCat API Platform as LongCat-2.5.
- Weights
- ProprietaryHosted access only
- Inputs
- textimage
- Context
- Not publishedLongCat API documentation lists 1,048,576 total input and output tokens per request. The available input budget depends on the requested output length.
- API price
- No per-token rateOnly limited-time promotional LongCat API prices are published: $0.30 per million uncached input tokens and $1.20 per million output tokens. Current standard hosted rates are not verified.
What LongCat says
- The new model has 1.6 trillion total parameters, with about 48 billion active per token.
- It adds native visual understanding and long-horizon coding and agent abilities to the LongCat-2.0 line.
Limitations
- Publisher API documentation gives a total request token limit, including output, rather than a fixed input-only context window.
- The published USD and CNY API token prices are limited-time promotional rates; standard hosted rates are not verified.
- A LongCat-2.5 checkpoint license and downloadable weights were not verified on November 5.
Publisher statements, not AI IQ measurements.
About this score
How much of LongCat 2.5's IQ is measured, and how the rest is estimated.
1/ 39benchmark results
1/ 7dimensions measured
6estimated
0/5Abstract
0/5Math
0/6Academic
0/6Coding
1/5Computer Use
0/6Reliability
0/6Knowledge Work
Bar height = share of benchmarks with resultsEstimatedNo score
Measured results are shown as published
Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.
Estimates fill the gaps
Abstract Reasoning, Mathematical Reasoning, Academic Reasoning, Programmatic Reasoning, Reliability and Knowledge Work have no published results yet, so their IQ is estimated from LongCat 2.5's other dimensions. Missing benchmarks can also draw on its predecessor, Longcat 2.0; a predecessor is an estimation input, not evidence that this model has the predecessor's measured results.
A missing result is not a zero
Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.
Compare LongCat 2.5
Side-by-side pages for the models it is most often weighed against.
#1 overall
fable-5.2
IQ144 39 higherEffective cost$73.69
Predecessor
LongCat 2.0
IQ101 4 lowerEffective cost$2.19
China (Other) #1
Ling 3.1 Flash
IQ120 15 higher
Closest IQ
Nemotron 3.5 Bolt 74B-A9B 24GB
IQ105 same IQ
Most similar profile
Muse Glimmer
IQ106 1 higherEffective cost$0.563
#2 overall
Gemini 4 Carbon
IQ142 37 higher
LongCat 2.5Comparison model
Frequently asked questions
Quick answers about LongCat 2.5, drawn from the same data as the rest of this page.
How AI IQ scores models →Get this data from the API →- What is the IQ of LongCat 2.5?
- LongCat 2.5 has an estimated IQ of 105, which ranks #100 out of 180 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
- What is LongCat 2.5 best at?
- There isn't enough published data to say yet. LongCat 2.5 has measured benchmark results in only 1 of the seven dimension AI IQ scores, and strengths are only compared once at least two dimensions are measured and ranked.
- What is the EQ of LongCat 2.5?
- LongCat 2.5 doesn't have an EQ score yet. EQ (emotional reasoning) is calculated from EQ-Bench 4 and AttuneBench, and neither has published a result for LongCat 2.5. AI IQ doesn't estimate EQ from other benchmarks, so the score appears once one of those results is published.
- Which benchmarks has LongCat 2.5 been tested on?
- LongCat 2.5 has published results on 1 of the 39 benchmarks behind AI IQ's Composite IQ, including BrowseComp.
- How much does LongCat 2.5 cost?
- AI IQ doesn't have a verified standard per-token API price for LongCat 2.5, so it can't calculate an effective cost. This is common for models offered only through subscriptions, restricted-access programs, or self-hosting.
- Is LongCat 2.5 open source?
- No — LongCat 2.5 is a proprietary model. Its weights are not publicly available, so it is used through hosted APIs and products.
- Who developed LongCat 2.5?
- LongCat 2.5 was developed by LongCat, based in China.
- When was LongCat 2.5 released?
- LongCat 2.5 was publicly released on November 2, 2026.
- How large is the context window of LongCat 2.5?
- AI IQ hasn't verified an input context limit for LongCat 2.5 from its developer's documentation, so none is listed rather than guessing.
- How fast is LongCat 2.5?
- AI IQ doesn't have independent speed measurements for LongCat 2.5 yet. Speed comes from independent tests of a model's hosted API, which aren't available for every model — especially new, restricted-access, or self-hosted ones.