qwen/qwen3.5-122b-a10b
The Qwen3.5 122B-A10B 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. In terms of...
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| SiliconFlowfp8 | $0.260 | $2.08 | 262K | 100% |
| Alibaba | $0.260 | $2.08 | 262K | 100% |
| AtlasCloudfp8 | $0.300 | $2.40 | 262K | 100% |
| Novitabf16 | $0.400 | $3.20 | 262K | 100% |
Qwen3.5-122B-A10B costs $0.260 per million input tokens and $2.08 per million output tokens via OpenRouter, making it 117th cheapest of 298 paid models.
Qwen3.5-122B-A10B scores 35.9 on the Artificial Analysis Intelligence Index, ranking 60th of 178 benchmarked models, with a GPQA Diamond score of 83%.
Qwen3.5-122B-A10B generates around 93 tokens per second with 787ms time-to-first-token (p50), the 73rd fastest tracked model.
Qwen3.5-122B-A10B supports a 262K-token context window and can output up to 262K tokens. It accepts text, image, video input.