Alibaba Qwenopen weightsQwen3.5 397B A17B pricing
$0.50 per 1M input tokens and $3.60 per 1M output tokens. In real work that is 0.58¢ per chat answer and 1.6¢ to summarize a 20-page PDF.
What Qwen3.5 397B A17B costs per task
Token counts estimated from a measured model of the same lab (Qwen3.8 Flash), rescaled by tokenizer. Reasoning tokens are billed as output.
| Task | Cost | Tokens used | Price rank | Basis |
|---|---|---|---|---|
| Answer a chat question | 0.58¢$5.80 / 1,000 | 88 in · 689 out · 909 reasoning | 237 of 336 | estimated |
| Reply to a customer email | 0.26¢$2.63 / 1,000 | 326 in · 98 out · 587 reasoning | 246 of 336 | estimated |
| Write a 1,500-word blog post | ≥5.8¢$57.65 / 1,000 | 104 in · 0 out · 16,000 reasoning | 302 of 336 | estimated |
| Write 10 product descriptions | ≥5.8¢$57.77 / 1,000 | 345 in · 0 out · 16,000 reasoning | 307 of 336 | estimated |
| Translate 1,000 words | 0.54¢$5.38 / 1,000 | 1,386 in · 1,166 out · 136 reasoning | 141 of 336 | estimated |
| Proofread one page (500 words) | 0.70¢$6.98 / 1,000 | 777 in · 671 out · 1,159 reasoning | 242 of 336 | estimated |
| Summarize a 20-page PDF | 1.6¢$16.07 / 1,000 | 13,119 in · 501 out · 2,141 reasoning | 228 of 336 | estimated |
| Extract data from an invoice | 0.81¢$8.09 / 1,000 | 501 in · 561 out · 1,617 reasoning | 267 of 336 | estimated |
| Analyze a 200-row spreadsheet | ≥6.1¢$60.78 / 1,000 | 6,357 in · 0 out · 16,000 reasoning | 253 of 336 | estimated |
| Answer from 5 retrieved documents (RAG) | 1.2¢$12.27 / 1,000 | 3,751 in · 974 out · 1,912 reasoning | 255 of 336 | estimated |
| Classify 100 support tickets | 3.1¢$31.02 / 1,000 | 1,463 in · 534 out · 7,879 reasoning | 292 of 336 | estimated |
| Review 500 lines of code | ≥6.0¢$60.04 / 1,000 | 4,874 in · 0 out · 16,000 reasoning | 262 of 336 | estimated |
| Write a SQL query | 0.96¢$9.56 / 1,000 | 177 in · 405 out · 2,225 reasoning | 273 of 336 | estimated |
| Run a 50-step coding agent | $1.04$1,040 / 1,000 | 2,259,875 in · 15,150 out · 80,850 reasoning | 238 of 336 | modeled |
Price at every provider
List prices per 1M tokens (input / output) at each host serving Qwen3.5 397B A17B through OpenRouter today. Hosts can differ in quantization, context and speed.
| Provider | Input / output | Cached input | Context | Uptime |
|---|---|---|---|---|
 Alibabaunknown | $0.39 / $2.34 | - | 262K | 100% |
 DeepInfrafp8 | $0.45 / $3.00 | $0.22 | 262K | 100% |
| D DigitalOceanunknown | $0.55 / $3.50 | $0.11 | 131K | 100% |
| P Phalaunknown | $0.55 / $3.50 | $0.23 | 262K | 100% |
| A AtlasCloudfp8 | $0.55 / $3.50 | $0.55 | 262K | 97% |
 Parasailfp8 | $0.50 / $3.60 | $0.30 | 262K | 100% |
| S StreamLakeunknown | $0.60 / $3.60 | $0.12 | 256K | 100% |
| G GMICloudfp8 | $0.60 / $3.60 | - | 262K | 100% |
 Novitaunknown | $0.60 / $3.60 | - | 262K | 100% |
| V Veniceunknown | $0.75 / $4.50 | - | 128K | 87% |
Price history
First seen in our price history on Feb 19, 2026 at $0.15 input / $1.00 output. Since then the output price moved from $1.00 to $3.60 per 1M tokens (10 changes).
FAQ
How much does Qwen3.5 397B A17B cost?
Qwen3.5 397B A17B costs $0.50 per 1 million input tokens and $3.60 per 1 million output tokens, with cached input at $0.30 (list price on Oct 18, 2026). In practice one chat answer costs about 0.58¢ and summarizing a 20-page PDF about 1.6¢.
Where is Qwen3.5 397B A17B cheapest?
Among 10 providers we track, the lowest list price today is at Alibaba: $0.39 input and $2.34 output per 1M tokens.
Is the Qwen3.5 397B A17B cost per task measured or estimated?
Estimated: Qwen3.5 397B A17B was not run directly. Tokens come from Qwen3.8 Flash and are rescaled by the tokenizer of Alibaba Qwen. Real usage can differ.