Key Info
OpenAI's GPT-6 Sol and Luna models are now available on the B.AI platform with 1.05M context, 128K max output, vision understanding, Computer Use, and 6-level adjustable reasoning. Sol targets complex software engineering while Luna offers a low-cost option for high-volume workloads.
Highlights
- GPT-6 Sol scores 68.8% on DeepSWE v1.1, reducing per-task costs by ~80%.
- GPT-6 Luna scores 66.6% on the same benchmark with input prices as low as $0.10 per 1M tokens.
- Both models support native vision understanding and Computer Use.
- Available via API and Web Chat simultaneously.