What's going on with Claude's new model release ('Fable') hitting usage limits so quickly?

Multiple members reported that a newly released Claude model (referred to as "Fable") burns through usage/session limits extremely fast — one member said it finishes a task in seconds but then hits the session limit almost immediately, forcing a switch back to a cheaper/older model. Another speculated this is tied to reports that the new release is intentionally "handicapped" — restricted from certain use cases like biomedical work and "frontier research" — which reportedly caused pushback from users who saw it as an anti-competitive move. One member noted that when they used it inside Claude Code that morning it was fast and more detail-oriented than before, without hitting a limit, though they guessed the task simply wasn't complex enough to trigger it. The group's takeaway: which model/plan to use depends heavily on the specific task.