LongCat 2.0
Language modelOpen source
LongCat 2.0
Meituan
- Released
- June 30, 2026
- Data date
- September 8, 2026
- Speed
- 48.5 tokens/sOutput throughput
- Price
- Input · Output / 1M tokens
- Input
- Text
- Output
- Text
LongCat 2.0 combines an N-gram embedding module with a mixture-of-experts component. Meituan lists 1.6 trillion total parameters, about 48 billion active per token, and a 1M-token context. The open model remains text-in and text-out.
Meituan released the weights under MIT in July 2026.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | Open weightsSource |
| Access | API, LocalSource |
| Parameters | 1.6TSource |
| Active parameters | 48BSource |
| Context window | 1,000,000 tokensSource |
| Notes | Open-weights MoE with 1.6T total parameters, about 48B active parameters, and a 1M-token context window.Source |
| License | MITSource |
| Release | 2026-06-30Source |
| Weights release | 2026-07-12Source |
| Parameters | 1,600 BSource |
| Active parameters | 48 BSource |
| Architecture | sparse MoE with N-gram Embedding (135B) and MOPD componentsSource |
| Training tokens | 35T+ (Meituan launch post summarizes the training corpus as more than 30T; the exact card's stronger >35T wording is retained.)Source |
| Input | TextSource |
| Output | TextSource |
| Tool calling | YesSource |
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
Terminal-Bench 2.1 (Show measurement, test conditions, and source)
- Source value
- 70.8
- Score
- 70.8
- Metric
- percent
- Unit
- %
- Category
- mixed
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Meituan LongCat 2.0 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor benchmark table; exact harness, snapshot and task settings are stated only where shown in the card.
SWE-Bench Pro (Show measurement, test conditions, and source)
- Source value
- 59.5
- Score
- 59.5
- Metric
- percent
- Unit
- %
- Category
- mixed
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Meituan LongCat 2.0 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor benchmark table; exact harness, snapshot and task settings are stated only where shown in the card.
SWE Multilingual (Show measurement, test conditions, and source)
- Source value
- 77.3
- Score
- 77.3
- Metric
- percent
- Unit
- %
- Category
- mixed
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Meituan LongCat 2.0 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor benchmark table; exact harness, snapshot and task settings are stated only where shown in the card.
FORTE (Show measurement, test conditions, and source)
- Source value
- 73.2
- Score
- 73.2
- Metric
- percent
- Unit
- %
- Category
- mixed
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Meituan LongCat 2.0 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Vendor benchmark table; exact harness, snapshot and task settings are stated only where shown in the card.
Output speed (Show measurement, test conditions, and source)
- Source value
- 48.4738057538424
- Score
- 48.4738057538424
- Metric
- tokens/s
- Unit
- tokens/s
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
Time per Intelligence Index Task (Show measurement, test conditions, and source)
- Source value
- 7.391913718629852
- Score
- 7.391913718629852
- Metric
- minutes
- Unit
- minutes
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
Latency reasoning time (Show measurement, test conditions, and source)
- Source value
- 41.259397088735184
- Score
- 41.259397088735184
- Metric
- seconds
- Unit
- seconds
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
Latency input time (Show measurement, test conditions, and source)
- Source value
- 3.92514578549998
- Score
- 3.92514578549998
- Metric
- seconds
- Unit
- seconds
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
Answer generation time (end-to-end timing breakdown) (Show measurement, test conditions, and source)
- Source value
- 10.314849272183796
- Score
- 10.314849272183796
- Metric
- seconds
- Unit
- seconds
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
End-to-end reasoning time component (Show measurement, test conditions, and source)
- Source value
- 41.259397088735184
- Score
- 41.259397088735184
- Metric
- seconds
- Unit
- seconds
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
End-to-end input time component (Show measurement, test conditions, and source)
- Source value
- 3.92514578549998
- Score
- 3.92514578549998
- Metric
- seconds
- Unit
- seconds
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Independent evaluation. unknown
- Context
- Artificial Analysis performance timing; output speed is decoded after the first streaming chunk. Provider basis: serving-provider median.
Our tests
SVG short test
One SVG prompt and fixed output settings. Captures are shown only with 0 reported reasoning tokens or documented lack of reasoning controls. Identical test conditions do not make the models' internal compute identical.
Earlier tests
Pelican on a bicycle
A community classic for free-form SVG drawing.
LongCat 2.0
Reasoning (requested): high
Current attempt
- Run date
- Oct 1, 2026
- API provider
- openrouter
- Requested API model
- meituan/longcat-2.0
- API model reported in response
- meituan/longcat-2.0
- Reasoning setting
- high
- Pinned endpoint
- AtlasCloud | meituan/longcat-2.0-20260720
- Temperature
- Provider default
- Token limit including reasoning
- 8,192 tokens
- Test protocol
- community-visual-tests-v1
- Technical checks
- Passed, not a quality rating
- Response time
- 47.1 s
Visually reviewed. This individual test does not establish overall model quality.
Prompt and test conditions
Generate an SVG of a pelican riding a bicycle
Token limit including reasoning: 8,192 tokens
Task origin (Simon Willison)Popularity
33.7 / 100
Sources and data date
| Type | Evidence and data date |
|---|---|
| Speed | Median across serving providers. Output speed after the first streaming chunk, not total response time. Retrieved September 8, 2026. |
| Price | Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. |
| Input | Documented modalities. Retrieved September 8, 2026. |
| Output | Documented modalities. Retrieved September 8, 2026. |
| Research status | Research date September 8, 2026. 4 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | knowledge-cutoff Original wording The exact LongCat 2.0 card and launch posts do not publish a training-data cutoff. |
| Additional source | Model metadata |
| Additional source | Model metadata (retrieved September 8, 2026) · Meituan LongCat 2.0 technical release · Editorial description reviewed October 2, 2026 |
| Additional source | Model metadata (retrieved September 8, 2026) · Meituan LongCat 2.0 open-weight release · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |