OpenAI and Anthropic Unveil Cheaper, More Powerful AI Models Amid Industry Debate
OpenAI and Anthropic both released new AI models on September 22, emphasizing lower costs and stronger performance, just as calls for slowing AI development gain traction in the tech world.
By Priya Raghunathan · First published 22 Sept 2026 · Chapter 2 of 2
Earlier: Anthropic Launches Claude Opus 5.5 AI Model Ahead of Planned IPO
In brief
- OpenAI has officially released GPT-6 Sol and GPT-6 Luna, offering significantly lower error rates and reduced operating costs.
- Anthropic launched Claude Opus 5.5, which now reportedly matches the performance of Fable 5.1 at a lower price.
- Both companies introduced their new models within hours of each other, intensifying the competition in AI pricing and performance.
- Industry leaders continue to call for a slowdown in AI development, but major labs are accelerating releases despite these concerns.
- Enterprises and developers now have access to high-performance AI models at roughly half the previous per-token cost.
Timeline · 5 moments
Anthropic launches Claude Opus 5.5 with lower cost and better performance
Digital Trends ↗OpenAI unveils GPT-6 Sol and Luna models after Anthropic's announcement
TechCrunch ↗Both companies highlight significant price cuts and improved efficiency in new models
The Next Web ↗Industry voices renewed concerns about the pace of AI development following model launches
NDTV Top Stories ↗Price war halves enterprise AI token costs
CIO ↗Update 23 Sept 2026, 3:34 pm UTC
OpenAI and Anthropic's new models have officially launched, with both companies now engaged in a price war that has halved per-token costs for enterprise buyers. The performance of Anthropic's Claude Opus 5.5 is now reported to approach that of Fable 5.1, while OpenAI's GPT-6 Sol and Luna models promise lower error rates and operational expenses.
How it started
Throughout 2026, the artificial intelligence sector has been marked by rapid innovation and rising concerns over the pace and direction of development. OpenAI and Anthropic, two of the most prominent players, have consistently introduced new models that push the boundaries of capability and affordability.
In the weeks leading up to September, industry leaders and even some company executives began voicing concerns about the unchecked speed of AI advancement. These calls for a slowdown were widely discussed, setting the stage for any new product launches to be scrutinized for both technical merit and broader implications.
How it unfolded
On September 22, Anthropic publicly announced Claude Opus 5.5, describing it as its most powerful and cost-effective model to date. The company emphasized that Opus 5.5 delivers performance comparable to its own Fable 5.1, but at a much lower price. Reports highlighted its enhanced capabilities in coding, knowledge work, and stricter safety features.
Just hours after Anthropic's launch, OpenAI revealed its own new models: GPT-6 Sol and GPT-6 Luna. These new additions to the GPT-6 family were presented as more affordable and efficient versions, building on the technology behind the flagship Astra model. OpenAI stated that Sol and Luna were developed to balance performance, speed, and cost, directly addressing users' demands for more accessible AI tools.
Both companies highlighted their models' reduced prices. For example, GPT-6 Sol and Luna reportedly offer up to fifty percent lower costs per token compared to previous models. Anthropic similarly touted a 40 percent reduction in API pricing with Opus 5.5, making advanced AI more accessible to businesses and developers.
The launches drew immediate attention because they came so soon after major public discussions about slowing AI progress. Observers noted the apparent contradiction between these calls for caution and the companies' continued push to expand their product lines.
Where it stands
As of now, both GPT-6 Sol and Luna, as well as Claude Opus 5.5, are available for integration in various products and services. Companies and developers can access these models at lower costs, with early reports noting improvements in speed and accuracy.
Despite the ongoing debate about the risks of rapid AI advancement, the market for high-performance, affordable models appears more competitive than ever. Both OpenAI and Anthropic are signaling that they intend to keep innovating, even as the industry grapples with questions about regulation and safety.
What to watch
Attention will now turn to the real-world impact of these models on businesses and developers, especially regarding cost savings and practical performance. There is also speculation about whether industry calls for a slowdown will gain more traction, or if the pace of new AI releases will continue unabated.


