Inference
EN
Alibaba

Qwen3.5 397B A17B

Qwen3.5 397B A17B is a model from Alibaba. It has 5 scored results in 4 categories. It has no index.

  • Open weights
  • Input: text · image · video
  • Released 2026-02
UnrankedEnough results for an index.
IndexAll categories, out of 100.UnrankedMedian 51.0
CoverageIndex weight that has results.60%
SpeedOutput tokens per second.37/sMedian 67/s
First tokenSeconds to the first output token.1.2 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.60Median $0.60
Output / 1MUS dollars per 1M output tokens.$3.60Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.1438±3 #83Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
qwen/qwen3.5-397b-a17b
Max outputMaximum output tokens per request.
64K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.12
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
9 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
SiliconFlowDeprecatedHost$0.39$2.34262K
ChutesHost$0.45$3.00262K
Deep InfraHost$0.45$3.00262K
DigitalOceanCloud$0.55$3.50131K
Novita AIHost$0.60$3.60262K
OpenRouterRouter$0.60$3.60262K
Together AIDeprecatedHost$0.60$3.60262K
ScalewayCloud$0.673$4.04256K
OVHcloudCloud$0.71$4.25262K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.1 counted57.8—
CodingCode writing and repair.None counted——
ReasoningLogic problems and puzzles.1 counted52.6—
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.1 counted56.4—
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.2 counted52.0—

Results

5 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
τ²-bench RetailAgentic84.4%2 / 23
τ²-bench TelecomAgentic97.8%3 / 18
τ²-bench AirlineAgentic81.5%5 / 22
GPQA diamond · Epoch AIKnowledge86.4% ±2.466 / 217
OTIS Mock AIME 2024-2025Math82.2% ±5.881 / 200
Mystery Game PuzzlesReasoning18.0% ±3.949 / 80
FrontierMath-Tiers-1-3-v2-PrivateMath31.2% ±2.769 / 90
τ²-bench BankingAgentic9.8%27 / 28
SAGEEducation85.5% ±1.815 / 29

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseSierra ResearchMITEpoch AICC BYmodels.devMIT

History

5 releases
Output / 1M$3.60 → $3.60
23 Oct 202612 Nov 2026

Questions

How much does Qwen3.5 397B A17B cost?
$0.60 per 1M input tokens, $3.60 per 1M output tokens.
How fast is Qwen3.5 397B A17B?
37 output tokens per second, 1.2 s to the first token.
What is the context window of Qwen3.5 397B A17B?
262K tokens.
Is Qwen3.5 397B A17B open source?
Yes. Its weights are published.
When was Qwen3.5 397B A17B released?
February 2026.

More from Alibaba

All models
Qwen3.5 Plus 2026-02-15UnrankedQwen3.5 397B (Reasoning)55.3Qwen3.5 397B53.8Qwen3 Max ThinkingUnrankedQwen3 Coder NextUnranked