OpenAI gpt-oss-120b API Prices Rise 305%: What Developers Need to Know

The Numbers

OpenAI has repriced gpt-oss-120b with a +305% increase across both input and output tokens:

This isn't a rounding error. Input cost has roughly quadrupled; output cost has more than tripled.

Does It Matter for Your Workload?

It depends heavily on your token ratio and volume. A pipeline generating 10M output tokens per month just went from $1.70 to $6.00 on output alone — before touching input costs. At scale, that delta compounds fast.

For low-volume or mostly-input-heavy workloads (retrieval, classification, short completions), the hit is softer but still real: input pricing is now four times what you may have budgeted.

If you locked in any cost estimates or signed off on a budget based on the old rates, revisit those numbers before your next billing cycle.

What to Do Next

• Audit your token split. High output-to-input ratios feel this increase most.
• Compare alternatives. Several models in a similar capability tier sit at lower price points right now. Check the gpt-oss-120b — live specs & price history page for a direct side-by-side with current competitors.
• Watch for tiered or cached pricing. If OpenAI introduces prompt caching or volume discounts for this model, the effective rate could shift again.

Bottom line: this is a meaningful cost event, not a minor adjustment. Build the new numbers into your projections today.

Ezra, Scout AI Team