Models/Qwen/Qwen3.5 397B A17B

Qwen3.5 397B A17B

The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...

01

Identity

262,144Context in
65,536Context out
2026-02-16 (OpenRouter created 1771223018)Released
OpenWeights
Canonical name
Qwen3.5 397B A17B
Aliases
qwen/qwen3.5-397b-a17b, Qwen/Qwen3.5-397B-A17B
Continuum hosted id
Not listed as a Continuum hosted id
OpenRouter slug
qwen/qwen3.5-397b-a17b
Hugging Face
Qwen/Qwen3.5-397B-A17B
Modalities
text, image, video
Open weights Not on Continuum host list Open

Context window on the 2026-08-19 OpenRouter row: 262,144 tokens.

Max completion tokens on that row: 65,536.

Input / output on that row: $0.39 / $2.34 per 1M tokens.

Architecture fields: text+image+video->text · Qwen3.

Hugging Face id on that row: Qwen/Qwen3.5-397B-A17B.

02

Should I use this for coding agents

The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.

When you want the 2026-02-16 catalog SKU, not a later rename.

DeepSWE

No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.

omitted, not invented as of 2026-08-13 source
Artificial Analysispublished integer
34
bench
Artificial Analysis
version
Intelligence Index
harness
published integer
metric
Intelligence Index
value
34
independent as of 2026-08-19 source
Vals Index

We did not open a vals.ai card for this identity. Chip omitted.

omitted, not invented as of 2026-08-19 source

Ranked view: Best coding models: the independent leaderboard puts this row and every other card that carries an independent coding score on one board, with the confidence intervals left visible.

Named benches, printed only

Copied from AA or vals HTML we opened. 0.0% placeholder bars are omitted. Do not average these into the hero tiles, and do not invent CI, $/task, or steps for Artificial Analysis.

BenchPrintedNoteSourceURLAs of
GDPval-AA v2 23.2% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
τ³-Banking 13.4% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
Terminal-Bench v2.1 51.3% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
SciCode 42% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
Humanity's Last Exam 29% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
GPQA Diamond 89.3% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
CritPt 1.7% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
AA-Omniscience Accuracy 30.8% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
AA-LCR 72.7% Printed on the AA Intelligence Evaluations grid for Qwen3.5 397B A17B (Reasoning). Not a DeepSWE chip. Artificial Analysis Qwen3.5 397B A17B source 2026-08-19
03

When not to use it

Trust this before you buy

  • When you need a Continuum-hosted SKU from this lab, use the hosted card on this hub instead of this catalog row.
  • $0.39 in / $2.34 out per 1M on the 2026-08-19 OpenRouter row. Context 262,144 in / 65,536 out.
  • Need a denser published sibling on this hub? Qwen3.8 Max is the card already on the cluster.
Get Plus See plans

Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.

04

Artifact and Hugging Face downloads

Artifact

SPDX / license
apache-2.0
Total params
HF safetensors.total: 403,397,928,944 stored tensors
Native precision
HF tensors BF16 + F32
Repo size
HF API usedStorage 806,809,049,334 bytes
HF created
2026-02-16T04:55:12.000Z
Official HF repo
Qwen/Qwen3.5-397B-A17B
HF downloads
270,105
HF likes
1,552

Official weights id on the 2026-08-19 OpenRouter row: Qwen/Qwen3.5-397B-A17B. Community GGUF is a quant, not a second Continuum card.

Official weights

Use this model on Hugging Face Files on Hugging Face

This is the model repo, not Download for Mac. Downloads and likes from the Hugging Face API on 2026-08-19.

Hub file tree

Safetensor shards
94
Shard bytes
806,796,241,352 bytes
Other file bytes
23,325,009 bytes
Tree file count
107
Hub usedStorage
806,809,049,334 bytes

usedStorage and the sum of .safetensors sizes can differ (LFS pointers, non-shard files, duplicate copies). Cited separately. Source: recursive tree API, 2026-08-19.

PathBytes
model.safetensors-00001-of-00094.safetensors8,589,934,736
model.safetensors-00002-of-00094.safetensors8,589,934,736
model.safetensors-00003-of-00094.safetensors8,589,934,736
model.safetensors-00004-of-00094.safetensors8,589,934,736
model.safetensors-00005-of-00094.safetensors8,589,934,736
model.safetensors-00006-of-00094.safetensors8,589,934,736
model.safetensors-00007-of-00094.safetensors8,589,934,736
model.safetensors-00008-of-00094.safetensors8,589,934,736
model.safetensors-00009-of-00094.safetensors8,589,934,736
model.safetensors-00010-of-00094.safetensors8,589,934,736
model.safetensors-00011-of-00094.safetensors8,589,934,736
model.safetensors-00012-of-00094.safetensors8,589,934,736
model.safetensors-00013-of-00094.safetensors8,589,934,736
model.safetensors-00014-of-00094.safetensors8,589,934,736
model.safetensors-00015-of-00094.safetensors8,589,934,736
model.safetensors-00016-of-00094.safetensors8,589,934,736
model.safetensors-00017-of-00094.safetensors8,589,934,736
model.safetensors-00018-of-00094.safetensors8,589,934,736
model.safetensors-00019-of-00094.safetensors8,589,934,736
model.safetensors-00020-of-00094.safetensors8,589,934,736
model.safetensors-00021-of-00094.safetensors8,589,934,736
model.safetensors-00022-of-00094.safetensors8,589,934,736
model.safetensors-00023-of-00094.safetensors8,589,934,736
model.safetensors-00024-of-00094.safetensors8,589,934,736
model.safetensors-00025-of-00094.safetensors8,589,934,736
model.safetensors-00026-of-00094.safetensors8,589,934,736
model.safetensors-00027-of-00094.safetensors8,589,934,736
model.safetensors-00028-of-00094.safetensors8,589,934,736
model.safetensors-00029-of-00094.safetensors8,589,934,736
model.safetensors-00030-of-00094.safetensors8,589,934,736
model.safetensors-00031-of-00094.safetensors8,589,934,736
model.safetensors-00032-of-00094.safetensors8,589,934,736
model.safetensors-00033-of-00094.safetensors8,589,934,736
model.safetensors-00034-of-00094.safetensors8,589,934,736
model.safetensors-00035-of-00094.safetensors8,589,934,736
model.safetensors-00036-of-00094.safetensors8,589,934,736
model.safetensors-00037-of-00094.safetensors8,589,934,736
model.safetensors-00038-of-00094.safetensors8,589,934,736
model.safetensors-00039-of-00094.safetensors8,589,934,736
model.safetensors-00040-of-00094.safetensors8,589,934,736
model.safetensors-00041-of-00094.safetensors8,589,934,736
model.safetensors-00042-of-00094.safetensors8,589,934,736
model.safetensors-00043-of-00094.safetensors8,589,934,736
model.safetensors-00044-of-00094.safetensors8,589,934,736
model.safetensors-00045-of-00094.safetensors8,589,934,736
model.safetensors-00046-of-00094.safetensors8,589,934,736
model.safetensors-00047-of-00094.safetensors8,589,934,736
model.safetensors-00048-of-00094.safetensors8,589,934,736
model.safetensors-00049-of-00094.safetensors8,589,934,736
model.safetensors-00050-of-00094.safetensors8,589,934,736
model.safetensors-00051-of-00094.safetensors8,589,934,736
model.safetensors-00052-of-00094.safetensors8,589,934,736
model.safetensors-00053-of-00094.safetensors8,589,934,736
model.safetensors-00054-of-00094.safetensors8,589,934,736
model.safetensors-00055-of-00094.safetensors8,589,934,736
model.safetensors-00056-of-00094.safetensors8,589,934,736
model.safetensors-00057-of-00094.safetensors8,589,934,736
model.safetensors-00058-of-00094.safetensors8,589,934,736
model.safetensors-00059-of-00094.safetensors8,589,934,736
model.safetensors-00060-of-00094.safetensors8,589,934,736
model.safetensors-00061-of-00094.safetensors8,589,934,864
model.safetensors-00062-of-00094.safetensors8,589,934,864
model.safetensors-00063-of-00094.safetensors8,589,934,864
model.safetensors-00064-of-00094.safetensors8,589,934,864
tokenizer.json12,807,982
vocab.json6,722,759
merges.txt3,353,259
model.safetensors.index.json303,440
README.md94,793
tokenizer_config.json16,710
LICENSE11,544
chat_template.jinja7,756
config.json4,177
.gitattributes1,570
preprocessor_config.json390
video_preprocessor_config.json385
generation_config.json244

Showing the first 64 of 94 safetensor shards, plus 13 other files. The snapshot was capped at 64 shards when the tree API was read on 2026-08-19, so the remaining 30 are on the Files tab, not on this page. Totals above are the full tree API sum, not this slice. Byte sizes are Hub tree API size fields, never invented.

05

Continuum serving

OpenRouter slug qwen/qwen3.5-397b-a17b on the 2026-08-19 catalog.

First party: https://qwen.ai/.

Not on the Continuum host list fetched 2026-08-19.

Continuum hosted id not on the host list
Qwen3.5 397B A17B first partyOpenRouter listing

We list first-party and OpenRouter, plus Continuum only when the live public allowlist named the id. This is not a 15-host routing table.

06

Economics

OpenRouter 2026-08-19: $0.39 in / $2.34 out per 1M tokens. Context 262,144 in / 65,536 out.

No separate internal-reasoning price on that row.

Price this model Opens the pricing calculator preloaded with Qwen3.5 397B A17B.
07

Provenance

ClaimSourceAs of
OpenRouter id qwen/qwen3.5-397b-a17b; context 262144; created 1771223018; $0.39 / $2.34 per 1MOpenRouter /api/v1/models2026-08-19
AA Intelligence Index 34; 9 printed Index benchesArtificial Analysis model page2026-08-19
OpenRouter id qwen/qwen3.5-397b-a17b, context 262,144OpenRouter /api/v1/models2026-08-19
HF downloads 270,105Hugging Face API Qwen/Qwen3.5-397B-A17B2026-08-19
HF safetensors.total: 403,397,928,944 stored tensors; HF API usedStorage 806,809,049,334 bytes; created 2026-02-16T04:55:12.000Z; HF tensors BF16 + F32Hugging Face API Qwen/Qwen3.5-397B-A17B2026-08-19
94 safetensor shards; shard bytes 806,796,241,352; tree files 107Hugging Face tree API Qwen/Qwen3.5-397B-A17B2026-08-19
08

Compare, FAQ, and Get Plus

No other card in this cluster shares a published DeepSWE official row we can put next to this one. We will not compare on vendor-blog numbers.

Continue through Qwen's model family with Qwen3 Max Thinking.

FAQ

Why does the lab blog disagree with DeepSWE or Scale?

Lab posts pick a harness, an effort, a split, and sometimes a private eval set. DeepSWE publishes the official mini-swe-agent row with cost, tokens, and steps. Scale SWE-bench Pro only counts when the public shared-harness board has a row we can fetch. A higher lab number is usually a different test, not a better one.

OpenAI-compatible call

No Continuum hosted id is verified for this model, so this page does not invent a curl target. Use the first-party API or the OpenRouter slug qwen/qwen3.5-397b-a17b.

Get Plus See plans

Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.