Qwen3.5-397B-A17B (MoE) Architecture Parameters, Benchmarks & VRAM
Largest Qwen3.5 open-weight flagship (397B total / 17B active, 512 experts) with native vision & video.
🧬 Qwen3.5-397B-A17B (MoE) Architecture & Tensor Specifications
Verified directly against HuggingFace (Qwen/Qwen3.5-397B-A17B) config.json & Model Card
1. Backbone & Dimensions
qwen3_5_moe- Total / Active Params
- 397B / 17B
- Hidden Size (d_model)
- 4,096
- MoE Expert Dim
- 1,024
- MoE Experts
- 512 Routed · Top-10 + 1 Shared
- Activation / Norm
- SiLU (SwiGLU) · ε=1e-6
- Native Tensor Dtype
- BF16
2. Attention & Layer Layout
HYBRID- Num Hidden Layers
- 60 Layers (15 Global KV)
- Hidden Layer Layout
- 60 Hybrid MoE Layers: 15 × [3 × Gated DeltaNet→MoE + 1 × Gated Attention→MoE]
- Attention Heads (Q / KV)
- 32 Q / 2 KV (16:1 GQA)
- Head Dimension
- 256
- Linear / KV Compression
- Only 15 layers × 2 KV heads allocate KV cache; extensible to 1M tokens via YaRN
- Vision Encoder / MTP
- ViT 27L · MTP 1L
3. Context, RoPE & Tokenizer
Tokenizer- Native Context Length
- 262,144 tokens
- Max Extended Context
- 1,010,000 tokens
- RoPE Configuration
- mRoPE (θ=10,000,000 · partial=0.25 · YaRN 4x→1M)
- Vocab Size
- 248,320 (Untied)
- Special Token IDs
- BOS: 248044 · EOS: [248046,248044] · PAD: 248044
📋 Qwen3.5-397B-A17B (MoE) Official API & System Card Specifications
Verified against official Alibaba Qwen developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- qwen3.5-397b-a17b
- Release Date
- 2026-02-23
- Weight License
- Open-Weight (API Spec)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- ✅ 原生支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 256,000 tokens (256K)
- Max Output Tokens
- 66,000 tokens
- Input Modalities
- Text + Image + Video
- Output Modalities
- Text
- Direct China Access
- ✅ Direct CN + RMB
- Deployment Mode
- Open-Weight + Cloud API
3. Official Pricing & Cache
CNY / 1M- Input / 1M Tokens
- ¥ 1.2
- Cached Input / 1M
- ¥ 0.24 -80%
- Output / 1M Tokens
- ¥ 7.19
- Tiered Pricing Notes
- Alibaba Bailian: Hosted on Alibaba Cloud Bailian (CN Direct)
🏆 Official Benchmark Results
Sourced from official System Card / Model Card & technical reports
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks.
C-Eval is a comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context.
PhD-level Google-proof science reasoning questions
All 30 problems from the 2026 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999.
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code.
Dual-control conversational tool-use benchmark
Berkeley Function Calling Leaderboard V4 (BFCL-V4) evaluates LLMs on their ability to accurately call functions and APIs, including simple, multiple, parallel, and nested function calls across diverse programming scenarios.
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
📊 Qwen3.5-397B-A17B (MoE) Full Benchmark Scorecard
37 public benchmarks · ranks among 154 models · synced 2026-11-20
💻 Coding 3 items
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
A multilingual benchmark for issue resolving in software engineering that covers Java, TypeScript, JavaScript, Go, Rust, C, and C++.
Terminal-Bench 2.0 is an updated benchmark for testing AI agents' tool use ability to operate a computer via terminal.
🤖 Agents & Tools 10 items
t2-bench is a benchmark for evaluating agentic tool use capabilities, measuring how well models can select, sequence, and utilize tools to solve complex tasks. It tests autonomous planning and execution in multi-step scenarios.
WideSearch is an agentic search benchmark that evaluates models' ability to perform broad, parallel search operations across multiple sources. It tests wide-coverage information retrieval and synthesis capabilities.
Berkeley Function Calling Leaderboard V4 (BFCL-V4) evaluates LLMs on their ability to accurately call functions and APIs, including simple, multiple, parallel, and nested function calls across diverse programming scenarios.
A high-difficulty benchmark purpose-built to comprehensively evaluate LLM agents on the Chinese web, consisting of 289 multi-hop questions spanning 11 diverse domains including Film & TV, Technology, Medicine, and History.
BrowseComp is a benchmark comprising 1,266 questions that challenge AI agents to persistently navigate the internet in search of hard-to-find, entangled information.
VITA-Bench evaluates AI agents on real-world virtual task automation, measuring their ability to complete complex multi-step tasks in simulated environments.
Seal-0 is a benchmark for evaluating agentic search capabilities, testing models' ability to navigate and retrieve information using tools.
MCP-Mark evaluates LLMs on their ability to use Model Context Protocol (MCP) tools effectively, testing tool discovery, selection, invocation, and result interpretation across diverse MCP server scenarios.
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
DeepPlanning evaluates LLMs on complex multi-step planning tasks requiring long-horizon reasoning, goal decomposition, and strategic decision-making.
📐 Math 11 items
An improved version of the MMLU benchmark featuring manually re-annotated questions to identify and correct errors in the original dataset. Provides more reliable evaluation metrics for language models by addressing dataset quality issues found in the original MMLU.
Harvard-MIT Mathematics Tournament 2025 - A prestigious student-organized mathematics competition for high school students featuring two tournaments (November 2025 at MIT and February 2026 at Harvard) with individual tests, team rounds, and guts rounds.
Harvard-MIT Mathematics Tournament 2025 - A prestigious student-organized mathematics competition for high school students featuring two tournaments (November 2025 at MIT and February 2026 at Harvard) with individual tests, team rounds, and guts rounds.
All 30 problems from the 2026 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999.
Multilingual Massive Multitask Language Understanding dataset released by OpenAI, featuring professionally translated MMLU test questions across 14 languages including Arabic, Bengali, German, Spanish, French, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, Swahili…
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks.
Extended version of MMLU-Pro providing additional challenging multiple-choice questions for evaluating language models across diverse academic and professional domains. Built on the foundation of the Massive Multitask Language Understanding benchmark framework.
IMO-AnswerBench is a benchmark for evaluating mathematical reasoning capabilities on International Mathematical Olympiad (IMO) problems, focusing on answer generation and verification.
Polymath is a challenging multi-modal mathematical reasoning benchmark designed to evaluate the general cognitive reasoning abilities of Multi-modal Large Language Models (MLLMs).
SuperGPQA is a comprehensive benchmark that evaluates large language models across 285 graduate-level academic disciplines.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions.
📜 Long Context 2 items
🧠 Reasoning & Science 5 items
C-Eval is a comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context.
Global PIQA is a multilingual commonsense reasoning benchmark that evaluates physical interaction knowledge across 100 languages and cultures.
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code.
MultiChallenge is a realistic multi-turn conversation evaluation benchmark that challenges frontier LLMs across four key categories: instruction retention (maintaining instructions throughout conversations), inference memory (recalling and connecting details from previous…
📚 Knowledge & General 6 items
Instruction-Following Evaluation (IFEval) benchmark for large language models, focusing on verifiable instructions with 25 types of instructions and around 500 prompts containing one or more verifiable constraints.
MAXIFE is a multilingual benchmark evaluating LLMs on instruction following and execution across multiple languages and cultural contexts.
INCLUDE is a multilingual language-understanding benchmark of 197,243 multiple-choice questions drawn from regional exams in 44 languages, built to measure regional and cultural knowledge rather than translated English content (Romanou et al., 2024).
WMT24++ is a comprehensive multilingual machine translation benchmark that expands the WMT24 dataset to cover 55 languages and dialects.
Instruction Following Benchmark evaluating model's ability to follow complex instructions.
NOVA-63 is a multilingual evaluation benchmark covering 63 languages, designed to assess LLM performance across diverse linguistic contexts and tasks.
Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).
⚡ Qwen3.5-397B-A17B (MoE) Interactive VRAM & Token Speed Calculator
Physical 512 GB · Load 53.7%
MoE: reads only 10.31 GB active weights/step
Enable Q8/Q4 KV quantization to scale batch concurrency
OLLAMA_CONTEXT_LENGTH=16384 ollama run qwen3.5:397b-a17bMac Studio M5 Ultra 512G Fits Qwen3.5-397B-A17B (MoE) · Cost & TCO
Approx. $9,499 (~$19/GB VRAM) · 190W TDP
📊 Qwen3.5-397B-A17B (MoE) Quantization VRAM Breakdown (16K Context)
| Quantization | Bits/Weight | Weights | KV Cache (16K) | Total VRAM | Quality Note |
|---|---|---|---|---|---|
| FP16 / BF16 | 16 bpw | 794 GB | 0.7 GB | 811.88 GB | Lossless Full Precision |
| FP8 / Q8_0 | 8.5 bpw | 421.81 GB | 0.7 GB | 432.25 GB | 99.9% Near-Lossless |
| GGUF Q6_K | 6.56 bpw | 325.54 GB | 0.7 GB | 334.05 GB | Ultra High Quality |
| GGUF Q5_K_M | 5.69 bpw | 282.37 GB | 0.7 GB | 290.02 GB | High Balanced Quality |
| Q4_K_M / AWQ | 4.85 bpw | 240.68 GB | 0.7 GB | 247.5 GB | Sweet Spot (Recommended) |
| GGUF IQ4_XS | 4.25 bpw | 210.91 GB | 0.7 GB | 217.13 GB | Compact 4-bit |
| GGUF Q3_K_M | 3.91 bpw | 194.03 GB | 0.7 GB | 199.92 GB | Moderate Perplexity Loss |
| GGUF Q2_K | 3.15 bpw | 156.32 GB | 0.7 GB | 161.45 GB | Extreme Compression |
🖥️ GPU & Apple Silicon Compatibility for Qwen3.5-397B-A17B (MoE) (Q4_K_M)
Frequently Asked Questions
What are the exact architecture parameters of Qwen3.5-397B-A17B (MoE)?
Qwen3.5-397B-A17B (MoE) uses Qwen3_5MoeForConditionalGeneration with 60 layers (60 Hybrid MoE Layers: 15 × [3 × Gated DeltaNet→MoE + 1 × Gated Attention→MoE]), hidden_size=4096, 32 Q heads / 2 KV heads (head_dim=256), and vocab_size=248,320.
How much VRAM does Qwen3.5-397B-A17B (MoE) require?
At 4-bit quantization (Q4_K_M, 16K context), Qwen3.5-397B-A17B (MoE) requires 247.5 GB VRAM. Full FP16 requires 811.88 GB.