HasItGot
Tech

Anthropic and OpenAI Launch Cheaper, More Capable AI Models

Despite talks of slowing down artificial intelligence progress, Anthropic and OpenAI have launched cheaper and more powerful new models.

HasItGot Newsroom
Anthropic and OpenAI Launch Cheaper, More Capable AI Models

Despite recent industry debates about slowing down the pace of artificial intelligence development, both Anthropic and OpenAI have launched a new wave of systems. According to a report by Engadget, these companies are introducing powerful updates to their existing platforms while simultaneously reducing usage costs for enterprise and consumer clients.

Anthropic's newly released model, called Opus 5.5, targets business users with enhanced capabilities for financial analysis and software debugging. Benchmarks provided by the company indicate that Opus 5.5 outperforms competing options in agentic coding tasks across tests like Terminal-Bench 4.0 and FrontierCode v1.1.

Better Performance and Lower Pricing

Alongside raw capability improvements, Anthropic states that Opus 5.5 generates clearer text and attempts to bypass safety boundaries roughly 85 percent less frequently than previous versions. The financial barrier to entry has also dropped, with input tokens priced at $4 and output tokens at $20, marking a decrease from the older Opus 5 pricing.

OpenAI has similarly concentrated on cost efficiency and accuracy through the rollout of its GPT-6 Sol and GPT-6 Luna models. These systems deliver professional work and computer use enhancements at rates up to 50 percent cheaper than previous promotional pricing tiers. Specifically, GPT-6 Sol charges $2 for input tokens and $10 for output tokens, while GPT-6 Luna costs $0.10 for inputs and $0.50 for outputs.

Accuracy and Safety Adjustments

In addition to cheaper pricing, OpenAI reports that GPT-6 Sol cuts factual errors in half compared to its predecessor and matches the coding performance of Fable 5.1. Both new OpenAI variants benefit from prompt caching improvements that increase response speeds, alongside better alignment measures that reduce false claims about coding results and block unsafe commands more reliably.

Both companies have also floated criteria for third-party evaluators to judge safety and progress, though it remains unclear if these suggestions will result in a unified industry standard. For users looking to access these tools, availability varies by platform and subscription level.

  • Anthropic Opus 5.5: Available to developers via Claude, Amazon Web Services, Google Cloud, and Microsoft Azure.
  • OpenAI GPT-6 Sol and Luna: Accessible in ChatGPT Work and Codex for Plus, Pro, Business, and Enterprise tiers.
  • GPT-6 Luna Only: Provided to free users and Go subscribers through the OpenAI desktop application.

More tech

All tech