Meituan LongCat-2.0 (1.78T MoE) Architecture Parameters, Benchmarks & VRAM
Meituan MIT-licensed 1.78T MoE supporting 1M context and 262K max output.
🧬 Meituan LongCat-2.0 (1.78T MoE) Architecture & Tensor Specifications
Verified directly against HuggingFace (meituan-longcat/LongCat-2.0) config.json & Model Card
1. Backbone & Dimensions
moe- Total / Active Params
- 1.776T / 48B
- Hidden Size (d_model)
- 8,192
- MoE Experts
- 768 Routed
- Activation / Norm
- SiLU (SwiGLU) · ε=0.00001
- Native Tensor Dtype
- BF16 / FP8
2. Attention & Layer Layout
MLA- Num Hidden Layers
- 38 Layers
- Hidden Layer Layout
- 38-Layer Sparse MoE (768 Experts)
- Attention Heads (Q / KV)
- 64 Q / 8 KV (8:1 GQA)
- Head Dimension
- 128
3. Context, RoPE & Tokenizer
Tokenizer- Native Context Length
- 262,144 tokens
- RoPE Configuration
- RoPE (原生 262,144 Context)
- Vocab Size
- 163,840 (Untied)
📋 Meituan LongCat-2.0 (1.78T MoE) Official API & System Card Specifications
Verified against official Meituan 美团 LongCat developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- meituan/longcat-2.0
- Release Date
- 2026-06-30
- Weight License
- Open-Weight (API Spec)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- 部分支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 1,024,000 tokens (1024K)
- Max Output Tokens
- 262,000 tokens
- Input Modalities
- Text
- Output Modalities
- Text
- Direct China Access
- ✅ Direct CN + RMB
- Deployment Mode
- Open-Weight + Cloud API
3. Official Pricing & Cache
USD / 1M- Input / 1M Tokens
- $ 0.3
- Cached Input / 1M
- $ 0.006 -98%
- Output / 1M Tokens
- $ 1.2
- Tiered Pricing Notes
- OpenRouter: Unified multi-provider routing via OpenRouter
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
⚡ Meituan LongCat-2.0 (1.78T MoE) Interactive VRAM & Token Speed Calculator
Physical 96 GB · Load 1168.6%
Exceeds usable VRAM by 1005.32 GB
MLA compresses multi-batch KV footprint
OLLAMA_CONTEXT_LENGTH=16384 ollama run meituan-longcat-2-0Insufficient Local VRAM? Ways to Run Meituan LongCat-2.0 (1.78T MoE)
Rent an hourly high-VRAM cloud GPU pod, or check the 12-month Buy vs Cloud vs API break-even curve.
📊 Meituan LongCat-2.0 (1.78T MoE) Quantization VRAM Breakdown (16K Context)
| Quantization | Bits/Weight | Weights | KV Cache (16K) | Total VRAM | Quality Note |
|---|---|---|---|---|---|
| FP16 / BF16 | 16 bpw | 3552 GB | 0.72 GB | 3624.21 GB | Lossless Full Precision |
| FP8 / Q8_0 | 8.5 bpw | 1887 GB | 0.72 GB | 1925.91 GB | 99.9% Near-Lossless |
| GGUF Q6_K | 6.56 bpw | 1456.32 GB | 0.72 GB | 1486.61 GB | Ultra High Quality |
| GGUF Q5_K_M | 5.69 bpw | 1263.18 GB | 0.72 GB | 1289.61 GB | High Balanced Quality |
| Q4_K_M / AWQ | 4.85 bpw | 1076.7 GB | 0.72 GB | 1099.4 GB | Sweet Spot (Recommended) |
| GGUF IQ4_XS | 4.25 bpw | 943.5 GB | 0.72 GB | 963.54 GB | Compact 4-bit |
| GGUF Q3_K_M | 3.91 bpw | 868.02 GB | 0.72 GB | 886.55 GB | Moderate Perplexity Loss |
| GGUF Q2_K | 3.15 bpw | 699.3 GB | 0.72 GB | 714.45 GB | Extreme Compression |
🖥️ GPU & Apple Silicon Compatibility for Meituan LongCat-2.0 (1.78T MoE) (Q4_K_M)
Frequently Asked Questions
What are the exact architecture parameters of Meituan LongCat-2.0 (1.78T MoE)?
Meituan LongCat-2.0 (1.78T MoE) uses LongcatCausalLM with 38 layers (38-Layer Sparse MoE (768 Experts)), hidden_size=8192, 64 Q heads / 8 KV heads (head_dim=128), and vocab_size=163,840.
How much VRAM does Meituan LongCat-2.0 (1.78T MoE) require?
At 4-bit quantization (Q4_K_M, 16K context), Meituan LongCat-2.0 (1.78T MoE) requires 1099.4 GB VRAM. Full FP16 requires 3624.21 GB.