OpenAI reduced the price of its GPT-5.6 Luna model by 80 per cent effective July 30. The new rates stand at $0.20 per million input tokens and $1.20 per million output tokens. Terra pricing also fell by 20 per cent to $2 and $12 respectively. Sol pricing remains unchanged. These adjustments apply across ChatGPT Work, Codex, and the OpenAI API. The company states that Luna delivers performance comparable to leading models from a year ago. A task costing a dollar previously now runs for about six cents on Luna. OpenAI claims this represents nearly nine times faster execution.
The reduction stems partly from internal efficiency gains. GPT-5.6 Sol reportedly optimised GPU software autonomously, cutting deployment costs by 20 per cent. Token generation improved by more than 15 per cent through speculative decoding. External market pressure likely influenced the decision as well. Low-cost Chinese providers have intensified competition across the sector. Microsoft is now promoting its own MAI models as cheaper alternatives. A sustained price war could impact revenue growth at frontier labs. These entities rely on balance sheets tied to massive infrastructure investments.
- Luna pricing drops to $0.20 per million input tokens
- Terra pricing falls to $2 per million input tokens
- Price cuts become effective on July 30
