OpenAI has introduced two new models, GPT-6 Sol and GPT-6 Luna, shortly after Anthropic released Claude Opus 5.5. These additions serve as faster and more affordable alternatives to the flagship GPT-6 Astra, which remains the company's most advanced model for complex tasks.
The company has reduced API costs by 50% compared to the promotional rates of GPT-5.6. GPT-6 Sol is priced at $2 per million input tokens and $10 per million output tokens. GPT-6 Luna is available at $0.10 for input and $0.50 for output per million tokens.
Performance testing on the AutomationBench benchmark showed GPT-6 Sol achieving a 33.2% pass rate at $0.27 per task. In comparison, Claude Opus 5 scored 26.9% while costing over 11 times more per task. On the Agents' Last Exam, Sol reached a 56.4% completion rate.
OpenAI introduced a reasoning effort setting that allows users to adjust computing power for increased accuracy. In OSWorld 2.0 tests for computer-based tasks, GPT-6 Sol achieved a 60.5% score, nearly matching Claude Opus 5 at approximately 80% lower cost.
Both new models are currently being rolled out gradually to ensure service stability. They are not yet available within the standard ChatGPT application.
