Qwen: Qwen3.5 397B A17B
by QwenQwen: Qwen3.5 397B A17B is a large language model from Qwen. It costs $0.550 per million input tokens and $3.50 per million output tokens. Its context window is 262K tokens.
- Input / 1M tokens
- $0.550
- Output / 1M tokens
- $3.50
- Cached input / 1M
- $0.225
- Context window
- 262K
On repeated prefixes
tokens
Who serves it cheapest
10 hosts serve Qwen3.5 397B A17B. Same weights, same API - the price difference is pure margin and routing.
| Provider | Input / 1M | Output / 1M | Context | Throughput | Uptime 24h |
|---|---|---|---|---|---|
| AlibabaCheapest | $0.390 | $2.34 | 262K | - | 100.0% |
| DeepInfrafp8 | $0.450 | $3.00 | 262K | - | 97.3% |
| Parasailfp8 | $0.500 | $3.60 | 262K | - | 99.9% |
| DigitalOcean | $0.550 | $3.50 | 131K | - | 98.0% |
| Phala | $0.550 | $3.50 | 262K | - | 99.5% |
| AtlasCloudfp8 | $0.550 | $3.50 | 262K | - | 97.2% |
| StreamLake | $0.600 | $3.60 | 256K | - | 93.6% |
| GMICloudfp8 | $0.600 | $3.60 | 262K | - | 94.6% |
| Novita | $0.600 | $3.60 | 262K | - | 99.7% |
| Venice | $0.750 | $4.50 | 128K | - | 95.1% |
The spread between Alibaba and Venice is 1.9× for identical weights. Quantization and context limits differ, so check both columns before switching.
Benchmarks
Independent scores published alongside the catalogue.
- Intelligence index
- 19.1
- Coding index
- 48.2
- Agentic index
- 10.6
About 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...
Specifications
| Model ID | qwen/qwen3.5-397b-a17b |
|---|---|
| Provider | Qwen |
| Context window | 262K tokens |
| Max output | 236K tokens |
| Input modalities | text, image, video |
| Output modalities | text |
| Knowledge cutoff | - |
| Open weights | Yes - Qwen/Qwen3.5-397B-A17B |
| Released | February 16, 2026 |
Cheaper alternatives
Models that cost less than Qwen3.5 397B A17B while keeping at least half its context window and every input modality it supports.
Frequently asked
How much does Qwen: Qwen3.5 397B A17B cost?
$0.550 per million input tokens and $3.50 per million output tokens. Cached input reads cost $0.225 per million tokens.
What is the context window of Qwen: Qwen3.5 397B A17B?
262K tokens, with up to 236K tokens of output per request.
Which provider serves Qwen: Qwen3.5 397B A17B cheapest?
Alibaba at $0.390 per million input tokens - 1.9× cheaper than Venice, the most expensive of the 10 hosts serving it.
Confirm against the source: Qwen official pricing.