Anthropic and OpenAI roll out cheaper models amid intensifying AI race

Anthropic and OpenAI have launched cheaper AI models, Opus 5.5, and GPT-6 Sol and Luna, as the US firms face growing price pressure from Chinese rivals and demand for lower-cost AI rises

Anthropic and OpenAI on Tuesday launched cheaper versions of their advanced artificial intelligence (AI) models, stepping up competition for business customers as companies increasingly look for lower-cost AI tools.

 

The launches came within hours of each other on Tuesday and followed recent calls from both companies to slow the pace of development of the most advanced AI systems. The releases also come as US AI labs face growing competition from lower-cost open-weight models, particularly from China.

 

Anthropic cuts cost of latest Claude model

 

Anthropic launched Claude Opus 5.5, the first model in its new Claude 5.5 family. The company said the model delivers performance comparable to its more powerful Fable 5.1 model on most work, while costing about 40 per cent less to run than its predecessor, Opus 5.

 

 

Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens. Anthropic said the model is designed to offer stronger performance while reducing the cost of running complex AI workloads.

 

The company plans to follow the release with Sonnet 5.5 and Haiku 5.5 in the coming weeks.

 

Fable 5.1, Anthropic's higher-end model, costs $10 per million input tokens and $50 per million output tokens. The company said Opus 5.5 can deliver a similar level of performance for most work at a lower cost.

 

OpenAI targets everyday AI work

 

OpenAI also released GPT-6 Sol and GPT-6 Luna, lower-cost models built on the technology behind its GPT-6 Astra model, which was launched earlier this month.

 

The company said Sol and Luna are designed to bring the capabilities of Astra to everyday work while offering faster performance and lower costs. Their API prices are 50 per cent lower than the promotional pricing for the previous GPT-5.6 models, with OpenAI citing improvements in caching and inference efficiency.

 

The two companies are increasingly competing not only on the capabilities of their models but also on the cost of using them.

 

Competition from cheaper open models

 

Anthropic and OpenAI are also facing competition from open-weight AI models that can be customised and run on users' own hardware.

 

Many of these models have emerged from Chinese AI companies and are putting pressure on US labs to offer more affordable products. The shift marks a change from the earlier focus of the AI race, when companies largely competed to build the most capable models.

 

Anthropic is also looking to expand its customer base as it prepares for a potential initial public offering (IPO). Investors are closely watching whether the company can maintain its rapid growth as competition increases and customers become more sensitive to AI costs, the Financial Times reported.

 

OpenAI, meanwhile, has been cutting prices and expanding its range of models as it seeks to attract more users and business customers.

 

First launches after calls for slower AI development

 

Opus 5.5 is Anthropic's first model release since Chief Executive Officer (CEO) Dario Amodei called for the industry to "pace the frontier" of AI development.

 

Amodei has argued that the development of AI capabilities should be paced so that safety measures can keep up. OpenAI CEO Sam Altman has also backed the call.

 

Despite the calls for a slower pace at the frontier, Tuesday's releases show that both companies continue to develop and deploy new models, particularly those aimed at broader and more cost-sensitive use.

 

Anthropic said Opus 5.5 underwent extensive alignment testing and evaluation by outside organisations before its release. The model also includes safeguards for high-risk areas, including cybersecurity and biology. OpenAI has also outlined areas for further work with independent evaluators, including training, safeguards, preparedness and the investigation of AI-related incidents.