Qwen: Qwen3.5-122B-A10B

Qwen3.5-122B-A10B is a multimodal model from Qwen that accepts text, image, and video inputs, with a context window and maximum completion length both reaching 262,144 tokens. It supports tool use and reasoning, making it suitable for agentic workflows and multi-step tasks. Structured output support is unconfirmed based on available data. At $0.26 per million input tokens and $2.08 per million output tokens, it sits at a moderate price point for a large model. Its blended benchmark score of 64.5 covers only 3 benchmarks, so treat that figure as preliminary rather than comprehensive. The standout number is an agentic score of 87.5, which makes it worth shortlisting specifically for tool-calling and agent-oriented use cases. Buyers who need long video or document context alongside agentic capability will find it a practical candidate, though broader benchmark coverage would give more confidence for general-purpose comparisons.