Advertisement

Anthropic and OpenAI announce cheaper and more disciplined AI models

 

Anthropic and OpenAI announce cheaper and more disciplined AI models

Although the debate in the artificial intelligence industry has largely focused on slowing down development, this does not appear to be preventing Anthropic and OpenAI from releasing new models.

Hot Topics

Both companies are working on developing their previous versions, Fable 5.1 and GPT-6 Astra, with new models that offer similar performance levels but at lower costs.

Anthropic says its new Opus 5.5 model is better at handling complex tasks, "finding and fixing software inefficiencies," and financial and business analysis—offers specifically tailored to its enterprise customers. The company claims that Opus 5.5 outperformed GPT-6 Astra in proxying on both Terminal-Bench 4.0 and FrontierCode v1.1 benchmarks, potentially making it more appealing to developers.

The improvements also extend to how the model works. Anthropic claims it produces easier-to-understand code and that it "attempted to circumvent boundaries about 85% less than previous models," suggesting it will disobey instructions less often. Most importantly, its cost per token is lower ($4 for input and $20 for output, compared to $5 for input and $25 for output in the Opus 5 model).

The new OpenAI models focus on efficiency, have been trained in methods similar to the GPT-6 Astra, and offer improvements in professional work, programming and computer use, but at a cost up to 50% lower than the promotional price of their GPT-5.6 counterparts.

The input tokens for GPT-6 Sol cost $2, and the output tokens cost $10, while for Luna the input cost is $0.10 and the output cost is $0.50.

In terms of improvements, both models are more factually accurate, with Sol making roughly half the number of errors as its predecessor. In terms of programming, the company says GPT-6 Sol can match the performance of Fable 5.1 at a lower cost. The models also communicate more clearly. Thanks to improved command caching, OpenAI says they reuse more context and respond faster. The company asserts that the new models are more compliant, lying less about their programming results and rejecting more attempts to bypass safeguards.

Anthropic's Opus 5.5 model is now available to developers via Cloud, Amazon Web Services, Google Cloud and Microsoft Azure.

GPT-6 Sol and GPT-6 Luna from OpenAI are available in ChatGPT Work and Codex for Plus, Pro, Business and Enterprise customers, while free users and Go subscribers will only get GPT-6 Luna in the desktop application.

 what is your opinion on this news ?do let us know in the comments on tha comment box

 if you liked this news article please share on friends & family & don't forget to follow on website & social media

Post a Comment

Previous Post Next Post
Advertisement
Advertisement
Advertisement