Summary
On August 21, 2026, OpenAI cut developer pricing for its frontier GPT-5.6 Sol model by more than 20% for at least three months (through November 21, 2026). Input token pricing drops from $5 to $4 per million, output from $30 to $20, and cached input from $0.50 to $0.40, applied to the pay-as-you-go API, Codex credits, and eligible ChatGPT Work plans. The cut leaves Sol priced below Anthropic's Claude Opus 5 on both input and output.
What changed
OpenAI reduced GPT-5.6 Sol pricing by more than 20% (input $5 to $4, output $30 to $20, cached input $0.50 to $0.40 per million tokens) as a promotional rate through at least November 21, 2026, covering the API, Codex credits, and eligible ChatGPT Work plans.
Why it matters
A steep, if temporary, cut on a frontier model lowers the cost floor for agentic and coding workloads that burn large token volumes, and directly targets Anthropic by pricing Sol below Claude Opus 5. The pass-through to Codex credits cheapens autonomous coding runs and quickly reached gateways like Vercel AI Gateway.
Evidence excerpt
Input went from $5 to $4 per million tokens (down 20%), output from $30 to $20 (down 33%) and cached input from $0.50 to $0.40 (down 20%)... The reduction applies to the pay-as-you-go API, to Codex credits and to eligible ChatGPT Work plans... Sol now costs less than Anthropic's Claude Opus 5 on both input and output.