OpenAI has introduced two new models, GPT-6 Sol and GPT-6 Luna, designed to distribute the benefits of frontier intelligence through greater cost efficiency. These models utilize similar training methods as GPT-6 Astra but focus on delivering faster, more affordable performance for everyday tasks and applications.
- API prices for Sol and Luna are reduced by 50% compared with their GPT-5.6 promotional pricing.
- On AutomationBench, GPT-6 Sol outperforms Claude Opus 5 at just 9% of the cost per task.
- GPT-6 Sol scores 56.4% on Agents’ Last Exam at approximately 60% lower cost than Claude Opus 5.
- On DeepSWE v1.1, GPT-6 Sol achieves 68.8%, within 1.1 percentage points of Claude Fable 5’s highest score at roughly 80% lower cost.
- Prompt caching improvements deliver higher cache hit rates by default, offering discounts of 90% on cached input-token reads.
The models are available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users, as well as GPT-6 Luna for free and Go users. These releases aim to make advanced AI practical for more complex work by combining strong capabilities with significantly lower operational costs.