OpenAI Launches GPT-6 Sol and Luna Models

OpenAI announced today the launch of two new language models that it describes as “cut from the same cloth as Astra.” The company said the models—named Astra‑Plus and Astra‑Pro—build on the same architectural foundations that underpin its earlier Astra release, incorporating the same transformer‑based framework and fine‑tuning techniques that have become a hallmark of OpenAI’s recent generative AI offerings.

According to the company’s statement, Astra‑Plus is optimized for general-purpose text generation and conversational tasks, while Astra‑Pro adds specialized capabilities for code synthesis, data analysis, and multilingual support. Both models are available immediately through OpenAI’s API, with pricing tiers that mirror those of the existing GPT‑4 family. OpenAI highlighted that the new models deliver comparable performance to Astra on standard benchmarks while offering lower latency and higher throughput for enterprise workloads.

Industry observers note that the introduction of Astra‑Plus and Astra‑Pro expands OpenAI’s portfolio of domain‑specific models, positioning the company to compete more directly with rivals that have begun to offer specialized AI services. The launch underscores OpenAI’s strategy of iterating on proven architectures to meet the growing demand for high‑performance, adaptable language models across a range of applications.