LLM Reference

Qwen3.5-397B-A17B

Released
2026-02-16
Last refreshed
2026-06-29
Status
Researched 89d ago
Open sourceCommercial use: permittedMultimodalCodingRAGAgentsLong contextVisionClassificationJSON / Tool use

Qwen3.5-397B-A17B is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and agents
  • Workloads that can use a 262k context window
  • Buyers comparing 4 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Family
Qwen3.5
Released
2026-02-16
Context
262k
Parameters
397B
Architecture
Mixture of Experts
Openness
Open source
License
Apache 2.0OSI-approvedCommercial use: permitted
Weights
Available
Code
Unknown
Created by
AlibabaAlibaba

AI research institute of Alibaba Group.

Hangzhou, Zhejiang, China
Founded 2017
Website
Pricing
Output / 1M
$2.34
Input / 1M
$0.390

Cheapest of 4 routes · Alibaba Cloud PAI-EAS

Providers(4)
OpenRouterTogether AIAlibaba Cloud PAI-EASNovita AI
View 4 provider routes

About

Alibaba's largest Qwen3.5 model, featuring a Mixture-of-Experts architecture with 397B total parameters and 17B active per token (using 512 total experts with 10 routed + 1 shared active). Supports 201 languages with a native 262K token context window extensible to 1M tokens via YaRN. Includes a thinking/reasoning mode, tool calling with MCP integration, and unified vision-language capabilities through early fusion training.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ C

2 relevant benchmarks in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Agents

Q/$ C

4 relevant benchmarks in the decision map.

Provider price ladder

Compare all 4

Compare API pricing across 4 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MRoute
Alibaba Cloud PAI-EAS$0.390$2.34
Serverless
OpenRouter$0.390$2.34
Serverless
Novita AI$0.600$3.60
Serverless
Together AI$0.600$3.60
Serverless

Available via routers & gateways(1)

OpenRouter

Hybrid

Unified hybrid gateway to 400+ models from 60+ providers via a single OpenAI-compatible API, with optional auto-routing that selects the best model per prompt.

PassthroughTogether AI

Capabilities

VisionMultimodalReasoningJSON / Tool useStructured Outputs

Benchmark peer barsfor Coding

SWE-bench VerifiedRank 39 of 81
Claude Fable 5
96.0
Claude Opus 5
96.0
Claude Mythos Preview
93.9
Claude Opus 4.8
88.6
Qwen3.5-397B-A17Bcurrent
76.2
LiveCodeBenchRank 20 of 56
DeepSeek V4 Pro
93.5
Gemini 3.1 Pro Preview
91.7
DeepSeek V4 Flash
91.6
Qwen3.7-Max
91.6
Qwen3.5-397B-A17Bcurrent
83.6

Benchmark scores(13)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
Google-Proof Q&A89.3diamondObserved 2026-04-18—Artificial Analysis
MMLU PRO87.8From official HuggingFace model card (accuracy)Observed 2026-06-07—Source
Massive Multi-discipline Multimodal Understanding85.0—Observed 2026-04-19—Source
Instruction-Following Evaluation92.6—Observed 2026-04-19—Source
BFCL72.9v4Observed 2026-04-19—Source
SWE-bench Verified76.2SWE-bench VerifiedObserved 2026-05-12—Source
AIME 202691.3AIME 2026 (accuracy)Observed 2026-06-07—Source
Berkeley Function Calling Leaderboard v372.9BFCL-V4, from official model card (accuracy)Observed 2026-06-07—Source
Humanity's Last Exam28.7HLE with CoT, no tools, from official model card (accuracy)Observed 2026-06-07—Source
LiveCodeBench83.6LiveCodeBench v6 (pass@1)Observed 2026-06-07—Source
MultiChallenge67.6Multi-Challenge leaderboard rank 2 of 28 (accuracy%)Observed 2026-06-07—Source
τ-bench86.7TAU2-Bench, from official model card (accuracy)Observed 2026-06-07—Source
Terminal-Bench 2.052.5Terminal-Bench 2.0 (accuracy%)Observed 2026-06-07—Source

Migration checks

No linked migration route is available for this model yet.

Compare Qwen3.5-397B-A17B with other models

Show all 53 popular comparisonssorted by 7-day search impressions
Qwen3.5-397B-A17B vs Claude Haiku 4.534Qwen3.5-397B-A17B vs GLM-5 Turbo30Qwen3.5-397B-A17B vs Llama 3 8B Instruct30Qwen3.5-397B-A17B vs GPT-4o29Qwen3.5-397B-A17B vs Claude 3.7 Sonnet27Qwen3.5-397B-A17B vs Mistral Medium 3.526Qwen3.5-397B-A17B vs DeepSeek V3.125Qwen3.5-397B-A17B vs Grok 424Qwen3.5-397B-A17B vs Mistral Large 221Qwen3.5-397B-A17B vs Qwen2.5-72B-Instruct20Qwen3.5-397B-A17B vs Llama 3 70B Instruct20Qwen3.5-397B-A17B vs GLM-4V 9B18Qwen3.5-397B-A17B vs Llama 3.2 11B Vision16Qwen3.5-397B-A17B vs Qwen2-VL-72B-Instruct15Qwen3.5-397B-A17B vs Gemma 7B Instruct15Qwen3.5-397B-A17B vs Grok 4.315Qwen3.5-397B-A17B vs Claude Opus 4.513Qwen3.5-397B-A17B vs o313Qwen3.5-397B-A17B vs DeepSeek V312Qwen3.5-397B-A17B vs Phi-3 Mini 4k12Qwen3.5-397B-A17B vs Llama 3.1 70B Instruct12Qwen3.5-397B-A17B vs Qwen3.5-Flash12Qwen3.5-397B-A17B vs Gemini 3 Pro12Qwen3.5-397B-A17B vs GPT-5.3-Codex-Spark12Qwen3.5-397B-A17B vs DeepSeek R111Qwen3.5-397B-A17B vs GPT-5.211Qwen3.5-397B-A17B vs GPT-5.510Qwen3.5-397B-A17B vs Phi 3.5 Vision Instruct9Qwen3.5-397B-A17B vs Gemini 2.5 Flash9Qwen3.5-397B-A17B vs DeepSeek R1 05288Qwen3.5-397B-A17B vs GPT Realtime 27Qwen3.5-397B-A17B vs gpt-realtime-1.57Qwen3.5-397B-A17B vs Gemini 3.1 Flash-Lite7Qwen3.5-397B-A17B vs Qwen3.5-35B-A3B6Qwen3.5-397B-A17B vs GPT-5.5-Cyber6Qwen3.5-397B-A17B vs Llama 3.2 1B Instruct6Qwen3.5-397B-A17B vs GPT-5.5 Instant6Qwen3.5-397B-A17B vs Llama 3.2 90B Vision5Qwen3.5-397B-A17B vs GPT Realtime Translate4Qwen3.5-397B-A17B vs Claude Mythos Preview4Qwen3.5-397B-A17B vs Llama 2 13B Chat4Qwen3.5-397B-A17B vs Trinity-Large-Thinking3Qwen3.5-397B-A17B vs o3 Mini3Qwen3.5-397B-A17B vs Mistral Medium 3 Instruct3Qwen3.5-397B-A17B vs GPT-5.4 Pro3Qwen3.5-397B-A17B vs Gemini 2.5 Pro3Qwen3.5-397B-A17B vs Grok-33Qwen3.5-397B-A17B vs gpt-realtime3Qwen3.5-397B-A17B vs Mistral Large 2 (2407)2Qwen3.5-397B-A17B vs Mistral Medium 31Qwen3.5-397B-A17B vs Phi-3 Silica1Qwen3.5-397B-A17B vs DeepSeek R1 Lite1Qwen3.5-397B-A17B vs Mixtral 8x7B0