OpenAI GPT-6.1 Sol's cached input drops 95%; DeepSWE improves by 6.4 percentage points
According to the Sept. 30 DevDay 2026 developer conference announcement, OpenAI released the new GPT-6.1 Sol model through its official X account and blog, positioning it as “the most cost-effective” mid-range flagship model available today. In terms of performance, official benchmark tests show that GPT-6.1 Sol improved by 6.4 percentage points over its predecessor, GPT-6 Sol, on DeepSWE v1.1. API Pricing: Cached Input Reduced to $0.10 (95% Cheaper Than Standard Input and 50% Lower Than Its Pre
Bisc·09-30 02:09
