Qwen3 Next 80B A3B Instruct Input Price Cut 10% to $0.09/1M Tokens

What Changed

Qwen has quietly cut the input price on Qwen3 Next 80B A3B Instruct by 10%, dropping from $0.10 to $0.09 per 1M tokens. Output pricing stays flat at $1.10 per 1M tokens.

Small move on paper — but for input-heavy workloads like document processing, RAG pipelines, or long-context summarization, the gap compounds quickly at scale.

Does It Matter for Your Workload?

The cut is input-side only, so its value depends entirely on your input/output ratio:

• Input-heavy tasks (large context, retrieval, classification): you'll see the full benefit of the -10% reduction.
• Output-heavy tasks (long-form generation, code synthesis): the $1.10/1M output price is unchanged, so this cut moves the needle less.

At $0.09 input and $1.10 output, Qwen3 Next 80B A3B Instruct remains competitively priced for a model of this scale — particularly given its sparse MoE architecture, which activates only 3B parameters per forward pass despite the 80B total parameter count.

Worth Benchmarking Against Alternatives

If you're already using this model, the math just got marginally better. If you're evaluating MoE options in this tier, the updated numbers are worth a fresh look. Check the Qwen3 Next 80B A3B Instruct — live specs & price history page for current pricing alongside comparable models on the site.

Bottom Line

Not a dramatic reprice, but a directionally positive signal. Qwen appears to be nudging this model toward broader adoption with incremental input cost reductions. Keep an eye on output pricing — that's where most cost sits for generative use cases.

Ezra, Scout AI Team