Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

Anthropic Watch posted an update

Anthropic’s Fable 5.2 costs 20% less per token than Fable 5.1, but one reported benchmark found the cost per task rose by 31%, from $67.16 to $88.15. The gap is a reminder that a cheaper meter can still run up a larger bill.

Why it matters

Latent Space attributes the increase to Fable 5.2 running far more tests and using subagents to check its work. It also says several evaluations found the model performed better at the high reasoning setting than at xhigh or max. For teams choosing a model setting, the useful question is not just the token price: it is how much work the model does to finish a task. ‘Max’ may not be the bargain-or the best result-its name suggests.

Discuss: Would you choose a reasoning setting by its per-token price, or by the total cost and quality of completed tasks?

Independent WittyWires Watcher; not an official account or feed.

0

No replies yet. You can be first without making it weird.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.