Qwen3.5 397B A17B Drops Input Price 33% to $0.39 per 1M Tokens

What Changed

Qwen has cut pricing on Qwen3.5 397B A17B across both token directions. Input drops from $0.55 to $0.39 per 1M tokens, and output falls from $3.50 to $2.34 per 1M tokens — a -33% reduction across the board.

Does It Move the Needle?

For read-heavy or retrieval-augmented workloads where input tokens dominate, the $0.39 input rate is genuinely competitive for a model at this parameter scale. The output cut matters more for generation-heavy tasks: at $2.34 per 1M output tokens, costs on a chatbot or long-form writing pipeline drop meaningfully at volume.

To illustrate: a workload burning 10M output tokens monthly just went from $35.00 to $23.40 — a $11.60 monthly saving before any input costs are counted.

Who Should Care

• High-volume API users routing through Qwen3.5 397B A17B will see automatic savings with no integration changes required.
• Teams evaluating MoE alternatives should reprice their benchmarks now — the gap between this model and some smaller, cheaper options has narrowed.
• Cost-sensitive prototypers who previously ruled out a 397B-class model on price may want to revisit the numbers.

Check current rates, context limits, and how this model stacks up against alternatives on the Qwen3.5 397B A17B — live specs & price history.

Bottom Line

A -33% price cut on both input and output is a real reduction, not a rounding adjustment. If you're running Qwen3.5 397B A17B at any meaningful scale, your bill just got smaller automatically. If you weren't — it's worth a second look.

Ezra, Scout AI Team