← sources

I tested Claude Fable 5.2 against Opus 5.5 on real tasks. The cost difference was interesting.

source: reddit (reddit:ClaudeAI)published 2026-11-15 11:16ingested 2026-11-16 02:35

retained text

**TL;DR — What I found:**

Fable 5.2 was much faster at building the 3D room (7 vs 20 minutes). Both models handled the warehouse simulation safely, but Opus was more efficient and cheaper in that run. Both were perfect across three receipt-processing attempts.

The bonus experiment was interesting: when I let them build an expense app without specifying the features, neither thought to handle duplicate receipts or refunds, even though both had dealt with those problems when explicitly asked.

**My conclusion:** Opus 5.5 remains my choice for everyday work. Fable's higher token price didn't consistently translate into better results in these tests. Of course, four experiments aren't enough to judge every use case.  
  
I've been curious whether Fable 5.2's higher token price actually translates into better performance than Opus 5.5.

So I put both models through the same challenges across a few different areas:

* **Creative development:** Turning a rough idea into something interactive.
* **Technical problem-solving:** Building and optimizing software under real constraints.
* **Everyday reasoning:** Working through messy information with specific rules.
* **Open-ended app building:** Giving the models a problem and letting them decide what the solution needs.

I wanted to look beyond benchmark scores and compare the actual outputs, reliability, time taken and cost.

What stood out to me was that the answer wasn't as straightforward as the token pricing might suggest.

For those who've used both models, where have you noticed a meaningful difference?

I recorded the experiments and outputs here, if you're interested: [https://youtu.be/E9ERgkouGmk](https://youtu.be/E9ERgkouGmk)

document fetched 2026-11-16 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xwt0ka/i_tested_claude_fable_52_against_opus_55_on_real