GPT-6 Sol vs Opus 5.5: 6 Surprising Things From the Same-Day Launch

On September 22, Anthropic released Claude Opus 5.5, and less than two hours later OpenAI released GPT-6 Sol and a smaller model, GPT-6 Luna. It's the second time this year the two companies have launched rival models on the same day, and this time both led with lower prices.

Anthropic says Opus 5.5 matches Claude Fable 5.1, its most capable public model, on most work and costs 40% less to run than Opus 5. OpenAI says Sol and Luna cost half as much as the GPT-5.6 models they replace.

"It's hard to overstate how competitive this pricing is," Simon Willison, an independent developer and co-creator of the Django web framework, wrote on his blog about OpenAI's new prices.

If GPT-6 Sol is the more affordable model and Opus 5.5 is smarter, the one you should choose depends on what you use AI for.

GPT-6 Astra: 10 Things to Know about ChatGPT’s New ModelOpenAI declared the AGI era, then shipped a model Sam Altman admits almost nobody can open.

/1. GPT-6 Sol Costs Half as Much as Claude Opus 5.5

Through the API, GPT-6 Sol costs $2 for every million tokens (chunks of text) it reads and $10 for every million it writes. Claude Opus 5.5 costs $4 and $20, twice as much, while GPT-6 Luna, OpenAI's budget model, costs $0.10 and $0.50.

Opus 5.5 is 20% cheaper than Opus 5, and Anthropic says it also uses fewer tokens, which is how it reaches its 40% saving. None of this changes subscriptions, though. If you pay for ChatGPT Plus or Claude Pro, your monthly bill stays the same.

/2. Claude Opus 5.5 Scores Higher Than GPT-6 Sol

Artificial Analysis, an independent firm that benchmarks AI models, put both through the same tests and combined the results into one overall score. At default settings, Opus 5.5 scored 51 and Sol 40. Even when set to think as hard as possible, Sol reached 48. Each test cost about $1.34 to run on Opus 5.5 and 25 cents on Sol.

Opus 5.5 writes code better, scoring 52.5% to Sol's 43.9% in the firm's coding test. But for everyday business tasks, there's almost nothing between them, with Sol at 61.7% and Opus 5.5 at 61.2%. If you code or work on hard problems, pick Opus 5.5. If your business runs lots of routine tasks and wants to spend less, Sol is what you should go for.

/3. What Claude Opus 5.5 Changes for Everyday Claude Users

If you pay for Claude, Opus 5.5 is available now on the Pro, Max, Team, and Enterprise plans, with higher five-hour usage limits. You also get one free "limit reset," which refills your usage allowance whenever you choose to use it. You can use it from Settings, then Usage, on the web or desktop app, but not on mobile.

Anthropic also says Opus 5.5 writes more clearly and with less jargon. Willison, who now uses it as his default model in Claude Code, said it "looks like it addresses the biggest complaints people had about Opus in terms of its communication style."