Anthropic released Claude Opus 5.5, described as matching Claude Fable 5.1 performance on most tasks while reducing inference costs by approximately 40 percent compared to Opus 5 (The Decoder). Independent benchmarking placed Opus 5.5 ahead of OpenAI’s GPT-6 Astra on most tasks despite the cost reduction (The Decoder).
OpenAI simultaneously launched GPT-6 Sol and Luna, two models that cut token pricing in half relative to previous generations while maintaining predecessor performance levels (The Decoder). Third-party analysis found minimal gains in raw intelligence metrics from the new OpenAI models (The Decoder). Both releases mark a shift toward token economics optimization rather than capability expansion.