Wednesday, 23 September 2026 | Updating Daily AI insight, written for builders

Claude Opus 5.5 and GPT-6 Sol: What Changes for Claude and ChatGPT Users

  • On 22 September 2026 both big AI companies launched new models within minutes of each other: Anthropic released Claude Opus 5.5, and OpenAI released GPT-6 Sol and GPT-6 Luna.
  • Claude Opus 5.5 is faster (over 30% faster than Opus 5), cheaper for developers, and Anthropic raised usage limits on its Pro, Max, Team and Enterprise plans.
  • GPT-6 Sol and Luna are cheaper, faster siblings of GPT-6 Astra. They are in ChatGPT Work and Codex for paying users; free users get Luna in the ChatGPT desktop app. They are not in regular ChatGPT chat yet.
  • If you only chat with ChatGPT on your phone for free, nothing changes for you today.

Tuesday, 22 September 2026 was an unusual day in AI: Anthropic and OpenAI each released new models within minutes of each other. If you use Claude or ChatGPT and just want to know what changed for you — without the benchmark tables — this is the short version.

Claude Opus 5.5: what’s new

Opus is Anthropic’s most capable everyday Claude model (its larger, more expensive model is called Fable). According to Anthropic, Opus 5.5:

  • Performs at the level of Claude Fable 5.1 on most work — Anthropic’s biggest, most expensive model — while costing much less to run.
  • Writes answers more than 30% faster than Opus 5.
  • Communicates more plainly. TechCrunch reports it uses less jargon and puts the important information at the start of its reply.
  • Scored best of any model Anthropic has tested on its automated behaviour audit, and was checked before release by outside evaluators including METR.

What it means if you use the Claude app

Anthropic says it is increasing the five-hour usage limits on the Pro, Max, Team and seat-based Enterprise plans alongside the launch — in plain terms, you can send more messages before hitting the “come back later” wall. Look for Opus 5.5 in the model menu at the top of a new chat. Anthropic’s announcement does not say whether Opus 5.5 reaches the free plan; check the model picker in your account.

Anthropic also said Claude Sonnet 5.5 and Claude Haiku 5.5 will follow “in the coming weeks” with many of the same improvements.

What it costs developers

For apps built on Claude, Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, down from $5 and $25 for Opus 5. (A token is a chunk of text about three-quarters of a word long.) You can compare it with every other model in our AI models database or estimate a monthly bill with the AI API cost calculator.

GPT-6 Sol and GPT-6 Luna: what’s new

Earlier in September OpenAI launched GPT-6 Astra, its most capable model. Sol and Luna are the smaller, cheaper members of the same family:

  • GPT-6 Sol is the middle option — for difficult work tasks, with higher usage limits than Astra.
  • GPT-6 Luna is the cheapest and fastest.

OpenAI highlights two things everyday users will notice:

  • Fewer factual mistakes. On OpenAI’s own test built from real conversations where users flagged errors, Sol makes about half as many mistakes as the model it replaces.
  • A clearer writing style: “more clarity, less jargon”, fewer odd turns of phrase and slightly shorter answers.

Who gets them in ChatGPT

Your ChatGPT plan What you get (from 22 September)
Free and Go GPT-6 Luna, in the ChatGPT desktop app
Plus, Pro Sol and Luna in ChatGPT Work and Codex
Business, Enterprise, Edu Sol and Luna in ChatGPT Work and Codex

One catch: OpenAI says the new models are “not yet available in Chat” — the ordinary ChatGPT conversation screen most people use. They are rolling out gradually, so if you don’t see them, try again later. Not sure which plan you are on or whether to upgrade? Our ChatGPT Free vs Go vs Plus vs Pro guide explains the differences.

What they cost developers

OpenAI halved the prices compared with the GPT-5.6 versions: GPT-6 Sol costs $2 input / $10 output per million tokens and GPT-6 Luna $0.10 / $0.50. Sol can take in about 1.05 million tokens at once, according to OpenAI’s developer documentation.

Claude Opus 5.5 vs GPT-6 Sol at a glance

Claude Opus 5.5 GPT-6 Sol
Company Anthropic OpenAI
Released 22 September 2026 22 September 2026
Position Top everyday Claude model Middle of the GPT-6 family (below Astra)
Main promise Fable-level results, 30%+ faster, cheaper Fewer mistakes, half the price of GPT-5.6 Sol
Developer price (input / output per 1M tokens) $4 / $20 $2 / $10
Where regular users find it Claude app model menu (paid plans get higher limits) ChatGPT Work and Codex (not yet in Chat)

They are not direct equivalents: Opus 5.5 is Anthropic’s top everyday model, while Sol sits under GPT-6 Astra. OpenAI’s announcement compares Sol with Claude models on tests OpenAI chose, and Anthropic publishes its own benchmark results — treat any company’s scores about itself as marketing until independent tests arrive.

So which should you use?

  • You use ChatGPT for free on your phone: nothing changes yet. Keep going.
  • You pay for ChatGPT Plus: try Sol in ChatGPT Work or Codex for longer or more complex tasks.
  • You pay for Claude Pro or Max: switch to Opus 5.5 in the model menu; the higher limits make it more practical than Opus 5 was.
  • You are choosing between the two apps: the model gap is small this month. Pick on features and feel — see our Claude vs ChatGPT comparison.

Whichever you use, the usual rule holds: AI models still make mistakes, including the new ones. Double-check anything important — health, money, legal questions — against a reliable source.

Frequently asked questions

When was Claude Opus 5.5 released?

On 22 September 2026. It is available in the Claude apps, through Anthropic’s API, and on Amazon Web Services, Google Cloud and Microsoft Azure.

Is Claude Opus 5.5 free?

Anthropic’s announcement talks about higher usage limits on the paid Pro, Max, Team and Enterprise plans and does not say whether free users get Opus 5.5. Check the model menu in your own Claude account.

Is GPT-6 available for free in ChatGPT?

Partly. Free and Go users can use GPT-6 Luna in the ChatGPT desktop app. GPT-6 Sol is for paying plans, and neither model is in the regular ChatGPT chat screen yet.

What is the difference between GPT-6 Astra, Sol and Luna?

Astra is OpenAI’s most capable and most expensive GPT-6 model. Sol is a faster, cheaper middle option for everyday hard work. Luna is the cheapest and fastest, aimed at simpler tasks and high volume.

Are Claude Sonnet 5.5 and Haiku 5.5 out?

Not yet. Anthropic says they will arrive in the coming weeks.

Written by Mustafa Ihsan

Mustafa Ihsan is the founder and editor of Convly.ai. He built and maintains the site's live AI models database, its price-performance index, and its free calculators for VRAM requirements, API costs and self-hosting economics. He writes about model pricing, benchmark results and the hardware needed to run AI models locally, and consistently prefers measured numbers to vendor claims.

Scroll to Top