Anthropic and OpenAI Launch Cost-Effective AI Models with Competitive Performance on September 22, 2026
Anthropic released Claude Opus 5.5 and OpenAI introduced GPT-6 Sol and GPT-6 Luna, both offering advanced AI capabilities at significantly reduced costs, underscoring a competitive push to improve AI affordability and performance.
On September 22, 2026, Anthropic unveiled Claude Opus 5.5, its newest AI model offering performance on par with its predecessor Claude Fable 5.1 but with a 40% reduction in operational costs. This cost efficiency is reflected in its pricing, set at $4 per million input tokens and $20 per million output tokens, marking a 20% discount compared to the previous Opus 5 model. This development indicates Anthropic's focus on balancing AI capability with economic scalability. [1]
In close timing, OpenAI launched two new AI models, GPT-6 Sol and GPT-6 Luna, which expand on the GPT-6 Astra architecture. GPT-6 Sol is targeted at handling complex tasks, while GPT-6 Luna is aimed at supporting high-volume, routine applications. Both models are offered at API prices roughly half those of GPT-5.6, with GPT-6 Sol priced at $2 per million input tokens and $10 per million output tokens, and GPT-6 Luna at exceptionally low costs of $0.10 per million input tokens and $0.50 per million output tokens. This strategic pricing may broaden access to sophisticated AI tools across varied business needs. [2]
These launches demonstrate a rapid pace of innovation in the AI sector, with Anthropic and OpenAI actively advancing model efficiency and affordability. Anthropic's earlier move on September 16 with the debut of Claude Docs also signaled an intent to elevate AI utility, particularly as it competes for partnerships such as those involving Microsoft. Meanwhile, OpenAI has concurrently acknowledged new challenges related to AI behaviors, committing to enhanced monitoring efforts, illustrating an ongoing balancing act between progress and responsibility in AI deployment. [3][4]