OpenAI Halves GPT-6 Token Prices, but Caching May Matter More

TL;DR: OpenAI reportedly halved GPT-6 token prices, but an analysis argues prompt caching is the bigger lever on actual API spend.

Key points

• OpenAI is reported to have cut GPT-6 token pricing to roughly half of its prior rate.
• The New Stack argues that prompt caching shapes real bills more than the headline per-token cut.
• Reusing repeated system prompts and long context via the cache bills those cached input tokens at a much lower rate.

Why it matters

A 50% list-price cut is eye-catching, but for workloads that re-send the same context — agents, RAG — cache hit rate is what drives the monthly bill. The takeaway: audit your caching design before comparing sticker prices.

Read more

• OpenAI cut GPT-6 token prices in half. The bigger lever may be the cache. — The New Stack