Qwen3 27B API Prices Jump 96% — Input Cost Nearly Doubles

What Changed

Qwen's 27B model has taken a sharp pricing step up. Input tokens now cost $0.42 per 1M tokens, rising from $0.214 — a +96% increase. Output tokens moved from $2.55 to $3.00 per 1M tokens. Both changes are live on the API.

Does It Matter for Your Workload?

The output increase (~18%) is manageable for most use cases. The input hike is the one to watch. Workloads that are read-heavy — document summarization, RAG pipelines, long-context classification — will feel this most. If you're processing millions of tokens on the input side daily, your costs roughly double overnight.

At $0.42 input / $3.00 output, Qwen3 27B is no longer the budget play it was. Developers who chose it specifically for cost efficiency should run the numbers against current alternatives before their next billing cycle.

What to Do Now

• Recalculate your monthly token spend using the new rates before assuming existing budgets hold.
• Check output-heavy vs. input-heavy ratios in your app — the impact is very asymmetric between these two changes.
• Compare alternatives on the site; several mid-size open-weight models sit in similar capability ranges at different price points.

See Qwen3.8 27B — live specs & price history for up-to-date pricing and side-by-side comparisons.

Bottom Line

This isn't a minor adjustment — a 96% input price increase is significant. It doesn't make the model bad, but it does change the value calculus for cost-sensitive applications. Worth a deliberate review rather than a quiet accept.

Ezra, Scout AI Team