Urgent.News

What's breaking now, across thousands of outlets.

AI

Claude Fable 5.2 vs Opus 5.5: which one should you actually use?

Anthropic released Claude Fable 5.2 on November 9, 2026. The first question everyone asks: should I switch from Claude Opus 5.5 ? This is the short version. The full guide, with Claude Code setup and a longer FAQ, is on my blog: Claude Fable 5.2 vs Opus 5.5: Which One Should You Use? Quick answer Opus 5.5 for most work. Coding, bug fixes, writing, data analysis. $4 / $20 per million tokens. Fable…

Anthropic recently introduced Claude Fable 5.2 on November 9, 2026. Users often wonder if they should switch from Claude Opus 5.5. A guide on this topic is available on the author's blog.

For most work, including coding, bug fixes, writing, and data analysis, Opus 5.5 is recommended at a cost of $4 for $20 per million tokens. Fable 5.2 is suggested for the most challenging and ambitious tasks, with a higher price of $8 for $40 per million tokens.

According to Anthropic's benchmark table, Fable 5.2 outperforms Opus 5.5 in several areas, especially for long, multi-step tasks. For example, in FrontierSWE v2 (agentic coding), Fable 5.2 scored 69.9% compared to Opus 5.5's 62.3%. Similarly, in Remote Labor Index (long knowledge work), Fable 5.2 scored 41.3% vs 21.3% for Opus 5.5. In Terminal-Bench-Science 0.1, Fable 5.2 achieved 74.6% while Opus 5.5 scored 58.7%. In OSWorld 2.1 (computer use), Fable 5.2 achieved 89.3% compared to Opus 5.5's 87.2%.

When considering the cost, if a task has 500K input tokens (with 400K cached) and 50K output tokens, Opus 5.5 would cost $1.48, whereas Fable 5.2 would cost $2.88. Running 20 such tasks daily for a month would result in $651 for Opus 5.5 and $1,267 for Fable 5.2. This significant cost difference is justified if Fable 5.2 proves to be accurate with complex tasks.

However, for routine work, Opus 5.5's lower cost makes it the preferred choice. It is important to note that Opus 5.5 in fast mode costs the same as Fable 5.2 at normal speed. Therefore, if speed is a priority, using Opus 5.5 in fast mode is a more cost-effective option.

The general guideline is to start with Opus 5.5 in default effort. If the results are unsatisfactory, adjust the effort to high, then xhigh. If the outcome is still not satisfactory, switch to Fable 5.2 for that particular task and revert back to Opus 5.5 once the challenging aspect is completed. To change models in Claude Code, update to Claude Code 2.1.321 or later by using the command "/model fable" for Fable 5.2 and "/model opus" for Opus 5.5.

To save the preferred model as the default for every new session, use "/model picker" followed by pressing Enter.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at dev.to →

More in AI

Restoring a Vertex AI Training Cluster: Recovering a Stopped GPU Node from Expired Reservations

Restoring a Vertex AI Training Cluster: Recovering a Stopped GPU Node from Expired Reservations Cluster Architecture Google Cloud’s Vertex AI Training Cluster (VTC) is designed for large-scale AI/ML…

  • Vertex AI Training Cluster uses GPU nodes for AI/ML training.
  • Expired reservation caused vtctest001-a4-0 node to TERMINATED state.
  • DWS Flex-Start model recovered node by preserving boot disk.

More from Thursday 12 November →