Large open Qwen multimodal MoE for visual agents and long technical tasksQwen3.5 397B-A17B by alibaba is offered by 41 providers on this page. Public prices are shown for 40 of them. Its official list price is $0.60 per 1M input tokens and $3.60 per 1M output tokens. The lowest paid channel is AIHubMix at $0.164 / $0.986 per 1M, about 3.6× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. It also has 1 free ($0) channel; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.
Median output speed is 58 tokens per second, with 2.12s to the first token.
The context window is 262,144 tokens at the reference host, but hosts report different limits, from 65,536 to 262,144, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, and Video. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2025-04.
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| UnoRouter qwen3.5-397b-a17b:free | Gateway | Free | — | — | 262,144 | 65,536 | deprecated | |
| AIHubMix | Gateway | $0.16 tiered >128K: $0.41 | $0.99 | — | — | 262,144 | 65,536 | |
| Alibaba (China)Official | First-party | $0.17 tiered >128K: $0.43 | $1.03 | — | — | 262,144 | 65,536 | |
| Merge Gateway | Gateway | $0.17 | $1.03 | $0.034 | — | 131,072 ⚠ | 32,768 | |
| EmpirioLabs AI qwen3-5-397b-a17b | Gateway | $0.17 tiered >128K: $0.43 | $1.03 | $0.17 | — | 258,048 ⚠ | 64,000 | deprecated |
| OrcaRouter | Gateway | $0.17 tiered >128K: $0.43 | $1.03 | — | — | 262,144 | 65,536 | |
| Qiniu | Gateway | $0.17 tiered >128K: $0.43 | $1.04 | — | — | 256,000 ⚠ | 64,000 | |
| SiliconFlow Qwen/Qwen3.5-397B-A17B | Cloud | $0.39 | $2.34 | — | — | 262,144 | 262,144 | deprecated |
| Kilo Gateway | Gateway | $0.39 | $2.34 | — | — | 262,144 | 81,920 | |
| TokenGo | Gateway | $0.40 | $2.65 | $0.20 | — | 262,144 | 65,536 |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 9 of 41 providers.
4 / 41 providers expose no reasoning control (reasoning_options: []).
1 more channels offer $0 (UnoRouter); free tiers usually have rate limits and no SLA, excluded from ranking.
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| SWE-Bench Verified | — | 76.4 | resolved | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and scale (percent correct, points or an index), so rows are not comparable with one another.