Qwen3.5 397B A17B API Pricing
Alibaba · Mid-tier · Qwen · released 2026-03. Prices in USD per 1 million tokens. Data snapshot: 2026-09-06.
- Input: $0.55 per 1M tokens
- Output: $3.50 per 1M tokens
- Batch API (input/output): not published
- Prompt-cache read: $0.23
- Context window: 262K
- Arena ELO: 1441 (as of 2026-09-01)
Cost per request by business use case
List price vs optimized (prompt caching on the cached input share, batch API for async workloads, only where the provider publishes those rates).
| Use case | Tokens (in/out) | Cost/req (list) | Cost/req (optimized) | Monthly @100K req |
|---|---|---|---|---|
| Support Ticket | 1,500 / 500 | $0.0026 | $0.0023 | $228 |
| Knowledge Q&A | 2,000 / 800 | $0.0039 | $0.0034 | $345 |
| Meeting Summary | 10,000 / 1,200 | $0.0097 | $0.0094 | $938 |
| Marketing Content | 2,500 / 1,800 | $0.0077 | $0.0075 | $751 |
| Coding Task | 3,000 / 2,000 | $0.0086 | $0.0082 | $816 |
| Invoice Processing | 1,500 / 600 | $0.0029 | $0.0028 | $278 |
| Call Summary | 2,000 / 700 | $0.0036 | $0.0035 | $349 |
| Agent Workflow | 6,000 / 3,000 | $0.014 | $0.012 | $1.2K |
Qwen3.5 397B A17B pricing by host
A host is a company that runs the model and sells it by the token — the lab that made it, a cloud platform like Amazon Bedrock, Azure or Vertex, or an independent provider. One model, several sellers, and often several prices. This model is sold by 10 different hosts. The price quoted above is the one we publish so that models stay comparable; it is not necessarily the cheapest way to buy this model. List prices as published on 2026-09-06 via OpenRouter — negotiated rates, committed-use discounts and provisioned capacity are not shown.
| Host | Input $/1M | Output $/1M |
|---|---|---|
| Alibaba | $0.39 | $2.34 |
| DeepInfra — fp8 | $0.45 | $3.00 |
| DigitalOcean (the price shown above) | $0.55 | $3.50 |
| Phala (the price shown above) | $0.55 | $3.50 |
| AtlasCloud (the price shown above) — fp8 | $0.55 | $3.50 |
| Parasail — fp8 | $0.50 | $3.60 |
| StreamLake | $0.60 | $3.60 |
| GMICloud — fp8 | $0.60 | $3.60 |
| Novita | $0.60 | $3.60 |
| Venice | $0.75 | $4.50 |
Models comparable to Qwen3.5 397B A17B
The nearest Mid-tier models on Arena ELO, so a substitution is compared on quality as well as price.
- GPT-5.1 (OpenAI) — $1.25 in / $10.00 out per 1M, Arena ELO 1439
- Qwen3.8 27B (Alibaba) — $0.42 in / $3.00 out per 1M, Arena ELO 1436
- GPT-5.2 (OpenAI) — $1.75 in / $14.00 out per 1M, Arena ELO 1436
- GPT-5 (OpenAI) — $1.25 in / $10.00 out per 1M, Arena ELO 1434
Common questions about Qwen3.5 397B A17B pricing
- How much does Qwen3.5 397B A17B cost?
- Qwen3.5 397B A17B costs $0.55 per million input tokens and $3.50 per million output tokens on the API.
- Does Qwen3.5 397B A17B offer batch or cached pricing?
- Prompt-cache reads cost $0.23/M input tokens. No batch pricing is published.
- What would Qwen3.5 397B A17B cost per month?
- For a support-ticket workload (1,500 input + 500 output tokens per request) at 100K requests/month, Qwen3.5 397B A17B costs about $258 at list prices.
These are list prices: real cost also depends on caching, batching, negotiated discounts, marketplace agreements, and existing contracts. Get a FinOps review from OptimNow or compare all models.
Build the business case for Qwen3.5 397B A17B in the AI ROI Calculator: add your business value on top of these costs to get ROI, break-even and payback.
All models · Compare side by side · Cloud compute pricing · Price barometer · FinOps guides · API and MCP server