OpenAI temporarily cuts GPT-5.6 Sol standard API rates
OpenAI announced the GPT-5.6 Sol rate cut on August 21, 2026, and made it effective the same day. The promotional rates are available at least through November 21, 2026. OpenAI has not said what pricing follows, so this should not be treated as a permanent cut.
The numbers below are standard short-context API rates per 1 million tokens. GPT-5.6 Sol longer-context tiers use different rates. Cache-write pricing follows OpenAI's 1.25x uncached-input multiplier.
| Rate | Before | Now | Change |
|---|---|---|---|
| Input | $5.00 | $4.00 | 20% lower |
| Cached input | $0.50 | $0.40 | 20% lower |
| Cache write | $6.25 | $5.00 | 20% lower |
| Output | $30.00 | $20.00 | 33% lower |
What the cut saves
- 1M input + 1M output: $35.00 before, $24.00 now, save $11.00 (31.4%).
- 10M input + 2M output: $110.00 before, $80.00 now, save $30.00 (27.3%).
- 1M input + 5M output: $155.00 before, $104.00 now, save $51.00 (32.9%).
Official and archival sources: OpenAI's dated API changelog and GPT-5.6 launch page establish the promotion and original rates; the current pricing table corroborates only the current side; an August 15 archived official pricing table preserves the pre-change rate card.