Skip to content

AI

Claude Opus 5.5 costs 40% less to run and still wants your job

Anthropic's Claude Opus 5.5 reportedly matches Claude Fable 5.1 on most work while costing 40% less to run than Opus 5. OpenAI dropped cheaper tiers the same day, because nobody wants to be the expensive robot.

Cover photo: Jernej Furman from Slovenia / CC BY 2.0

TL;DR

  1. Claude Opus 5.5 costs around 40% less to run than Opus 5, per Anthropic.
  2. Anthropic says it performs at Claude Fable 5.1 level on most work.
  3. OpenAI launched cheaper GPT-6 tiers the same Tuesday. The price war is officially on.

What happened

Anthropic released Claude Opus 5.5 on Tuesday, the first model in its new Claude 5.5 family. Anthropic claims it matches Claude Fable 5.1 on most tasks and runs 40% cheaper than Claude Opus 5.

The pitch is efficiency. Dianne Penn, Anthropic's head of product management, research and labs, told CNBC the company keeps working on making the model's thinking and answering use fewer tokens, depending on your effort setting.

5 the new leading model and a major step up from Opus 5. Its examples are the usual flex, but with numbers.

One early tester moved 680,000 lines of code to a new setup in under a day. Anthropic says a team of engineers would have needed weeks.

In another test, Anthropic told it to make every page of a web app load faster. Opus 5.5 succeeded 39 of 40 times. Opus 5 made smaller gains that also changed the app's behavior, which is the polite way of saying it broke things.

This is Anthropic's first release since it called for pacing the frontier. External evaluators, including Frontier Design and METR, tested the model before launch. 5 is the strongest model it has tested so far.

That is a vendor result, so file it accordingly.

Why it matters

Anthropic was not alone on Tuesday. OpenAI added GPT-6 Sol and GPT-6 Luna to its GPT-6 family, with API prices 50% below GPT-5.6 promotional pricing. Sol targets complex work like coding. Luna targets high-volume jobs such as extracting information or summarizing documents.

CNBC ties both launches to the same pressure: cheaper open-weight competitors. Two frontier labs cutting prices on the same day is not a coincidence. It is a market talking.

Frontier quality is getting cheaper, and the labs know open-weight rivals are the reason.

Note the order of operations. Anthropic called for pacing the frontier, and its first move afterward is a model that is cheaper and, it says, stronger. The slowdown talk and the shipping schedule are doing different jobs.

For anyone building on these APIs, the practical read is simple. If Opus 5.5 really matches Fable 5.1 on most work at a lower running cost, your bill and your model choice both deserve a second look.

What to do with it

  • Rerun your costs.Price your heaviest Opus 5 workloads against Opus 5.5 before your next invoice does it for you.
  • Test on your own tasks.Vendor benchmarks are slides, your codebase is the exam, so run the same prompts on both models.
  • Check the effort setting.Anthropic says token use varies with it, so tune it instead of leaving everything on maximum.
  • Compare the rivals.OpenAI's GPT-6 Sol and Luna landed the same day at 50% off GPT-5.6 promotional pricing, so get quotes from both.

Ciao, sapiens. The robots just got cheaper to hire, so brush up that résumé.