OpenAI Cuts GPT-5.6 Sol API and Codex Credit Pricing by 20%
OpenAI has lowered API and credit pricing for its frontier GPT-5.6 Sol model by more than 20% across API endpoints and eligible ChatGPT Work plans. The promotional discount is active through at least November 21, 2026.
Impact: Medium
Why it matters
Recalculate your token budgets and automated Codex workloads to leverage reduced frontier inference costs over the next three months.
TL;DR
- 01GPT-5.6 Sol API token and Codex credit costs are reduced by over 20% through November 21, 2026.
- 02Plus, Pro, and Business fixed subscription tiers remain unchanged.
- 03High-throughput agentic workflows and batch jobs can utilize Sol at lower marginal expense.
Key facts
- Price Cut
- >20% reduction on GPT-5.6 Sol
- Promotional Window
- August 21 – November 21, 2026
- Eligible Tiers
- API, Codex credits, ChatGPT Work
Frontier Model Price Reduction
OpenAI has introduced an immediate price drop of over 20% for GPT-5.6 Sol across developer API endpoints and Codex credit allocations. The promotional rates will remain available at least through November 21, 2026.
Impact on Work and Codex Tiers
The pricing reduction applies to:
- Developer API consumption per million tokens for
GPT-5.6 Sol - Codex credit consumption in automated coding pipelines
- Eligible ChatGPT Work enterprise workspace tiers
Standard subscriptions (Plus, Pro, and Business) maintain their current rate caps and pricing.
Model Cost Hierarchy
The Sol reduction follows earlier cuts across the GPT-5.6 family, where Terra was reduced by 20% and Luna by 80%. This brings frontier-level capability closer to high-volume development workflows without requiring downgrades to smaller models.
✓ When to use
- Running high-token agentic software engineering and code review workflows in Codex.
- Large-scale reasoning pipelines deployed via the OpenAI developer API.
✕ When NOT to use
- Workloads strictly requiring static, permanently locked-in API unit economics beyond November 2026.
- Tasks that are already cost-optimal on lightweight models like Luna.
What to do today
- Audit monthly token spend across active OpenAI API keys running GPT-5.6 Sol.
- Evaluate re-routing batch agent workflows from smaller models back to Sol during the discount window.
What the community says
“The price war is on, as expected. This is the true China ev moment for llm wars”
“Margin collapse”
Sources