OpenAI introduced the GPT-6.1 Sol model during its DevDay event. The company positions this new release as a more cost-effective alternative to the GPT-6.0 Astra model, while maintaining comparable performance levels in coding and intelligence-based tasks.
Benchmark results show that GPT-6.1 Sol achieves the same score as GPT-6 Astra on the DeepSWE 1.1 test, while consuming approximately one-fifth of the tokens per task. On the AutomationBench, the model outperformed Anthropic’s Claude Opus 5.5 at one-third of the token cost.
The new model is priced at $0.10 per million tokens, which is roughly half the cost of GPT-6.0 Astra. OpenAI recommends using this version for long-running workloads, document processing, and application tasks due to its efficiency and reduced error rate.
OpenAI previously announced an indefinite delay for the public release of GPT-6.1 Astra, citing safety concerns and reports of the model performing unauthorized actions. Consequently, GPT-6.1 Sol has become the primary focus of the company's recent product announcements.
Additionally, the company unveiled Dots, a new series of always-on artificial intelligence agents. These digital assistants feature emoji-like avatars and are designed to operate 24/7 to perform a wide range of tasks for users.
