Qwen 3.5 397B a17b API

Alibaba's Qwen 3.5 397B a17b, through an OpenAI-compatible API. 3 providers quote it now, and each request goes to the offer with the lowest estimated charge that meets your rules.

Providers compete for each request. The exchange ranks the offers that meet your rules by estimated charge. It fixes your maximum charge before the first token, and you pay for the tokens that the model uses.
The OpenAI and Anthropic APIs. Use the OpenAI or Anthropic SDK that you have. Set the base URL and your key, and send qwen3.5-397b-a17b as the model.
Your data rules. Send x-liquid-retention: none and only offers whose provider does not keep your content compete for the request.

Facts

Model name: qwen3.5-397b-a17b
Maker: Alibaba
Context window: 32K – 131.1K tokens
Input price: $0.168 to $0.500 per million tokens
Output price: $1.032 to $3.600 per million tokens
Providers: 3 independent providers
Regions: eu, global, us

Providers

EmpirioLabs AI: global, $0.172 input and $1.032 output per million tokens, $0.172 cache read per million tokens, 32K tokens of context
SOKKAN: eu, $0.168 input and $1.158 output per million tokens, $0.168 cache read per million tokens, 131.1K tokens of context
Pera: us, $0.500 input and $3.600 output per million tokens, $0.500 cache read per million tokens, $0.00048 fee for each request, 128K tokens of context

Scores

HLE: 29.0, #85 of 273
GPQA: 89.3, #53 of 250
AA-LCR: 77.3, #76 of 237
SciCode: 44.8, #81 of 123
Terminal-Bench Hard: 40.9, #26 of 166
τ²-Bench: 95.6, #14 of 169
IFBench: 78.8, #5 of 178
SimpleQA Verified: 20.3, #93 of 126

Scores: Artificial Analysis (https://artificialanalysis.ai/)

Scores: Epoch AI, 'Capabilities & Benchmarking' (https://epoch.ai/benchmarks)

Request

curl https://router.inference.ai.exchange/v1/chat/completions \
  -H "Authorization: Bearer $LIQUID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.5-397b-a17b",
    "messages": [{"role":"user","content":"Say hello."}]
  }'

Questions

What is Qwen 3.5 397B a17b?: Served by intent-router: each job goes to the cheapest healthy upstream for this model.
How much does Qwen 3.5 397B a17b cost?: Offers now ask $0.168 to $0.500 per million input tokens and $1.032 to $3.600 per million output tokens. A job pays for the tokens it uses. The exchange picks the cheapest eligible offer.
What is the context length of Qwen 3.5 397B a17b?: The largest offer takes 131,072 input tokens. A smaller offer can take less. The exchange skips an offer that cannot take your message.
What can Qwen 3.5 397B a17b do?: Its providers list Vision, Reasoning budget, Reasoning effort, JSON object, JSON schema, Frequency penalty, Log probabilities, Presence penalty, Repetition penalty, Seed, Stop sequences, Top-k sampling, Top-p sampling, Text, Tools.
What inputs and outputs does Qwen 3.5 397B a17b support?: Text · Images → Text.
Which providers serve Qwen 3.5 397B a17b?: 3 independent providers quote it now, with 3 offers. The table above ranks them by the cost of a typical job.
How do I call Qwen 3.5 397B a17b?: Send "qwen3.5-397b-a17b" as the model in an OpenAI or Anthropic request. The API section links to setup steps for each client.
When was Qwen 3.5 397B a17b listed?: The exchange first listed it on 10/8/2026. This is not its release date.