Qwen3.5-27B API Prices Rise 67%: Input and Output Costs Both Jump

What Changed

Qwen's 27B model has repriced at a meaningful premium. Input now costs $0.26 per 1M tokens, up from $0.195, and output has moved from $1.56 to $2.60 per 1M tokens — both representing a +67% increase.

Does It Matter for Your Workload?

For low-volume or exploratory use, this won't sting much. But output-heavy workloads — long-form generation, multi-turn agents, document drafting — will feel the difference immediately. At the new output rate of $2.60/1M tokens, a pipeline generating 10M output tokens monthly now costs $26 instead of $15.60. That's real budget impact at scale.

Input-heavy workflows (classification, retrieval-augmented generation with short answers) are less affected, since the input jump from $0.195 to $0.26 is proportionally the same but smaller in absolute dollar terms.

What To Do

If Qwen3.5-27B is a core part of your stack, now is a good time to audit your token split. High output ratios make this repricing hurt more. It's also worth benchmarking whether a smaller or competing model hits your quality bar at a lower price point — the 67% jump is large enough to justify the comparison exercise.

Check the current specs and track further price movements on the Qwen3.5-27B — live specs & price history page.

Bottom Line

This isn't a minor rounding adjustment. A 67% price increase on both input and output is a significant repricing for a mid-size model. Keep an eye on whether the cost-to-capability ratio still holds up against alternatives.

Ezra, Scout AI Team