Model lifecycle

Claude Fable 5

Released 2026-06-09 · Back to Claude

Exact-version signals observed 2026-06-12–2026-07-13 · n=8,995

Trend history still collecting The current window is ready; 2 of 4 independent segments are available.

Lifecycle updated daily · latest complete UTC day 2026-07-13

Fixed 14-day baseline
67.9
n=2,166
Latest 28 days
53.0
n=7,401
Change vs baseline
−15.0
90-day slope
pts / 30 days

14-day rolling experience

Each daily point summarizes the previous 14 complete UTC days. The dashed line is the fixed release baseline; the verdict still uses independent non-overlapping 14-day segments.

14-day rolling experience from 2026-06-23 to 2026-07-14; latest score 50.7, fixed baseline 67.9.4050607080Fixed baseline 67.92026-06-09–2026-06-23 · 67.9 · n=21662026-06-10–2026-06-24 · 68.1 · n=21952026-06-11–2026-06-25 · 68.0 · n=22362026-06-12–2026-06-26 · 67.9 · n=22602026-06-13–2026-06-27 · 70.5 · n=19302026-06-14–2026-06-28 · 72.9 · n=10352026-06-15–2026-06-29 · 73.2 · n=8862026-06-16–2026-06-30 · 71.9 · n=7682026-06-17–2026-07-01 · 67.2 · n=7442026-06-18–2026-07-02 · 50.9 · n=14422026-06-19–2026-07-03 · 50.7 · n=24292026-06-20–2026-07-04 · 50.0 · n=29812026-06-21–2026-07-05 · 50.1 · n=33292026-06-22–2026-07-06 · 50.9 · n=36142026-06-23–2026-07-07 · 51.1 · n=39952026-06-24–2026-07-08 · 50.5 · n=44542026-06-25–2026-07-09 · 51.1 · n=48482026-06-26–2026-07-10 · 51.2 · n=50872026-06-27–2026-07-11 · 51.1 · n=54822026-06-28–2026-07-12 · 51.0 · n=57192026-06-29–2026-07-13 · 50.4 · n=61792026-06-30–2026-07-14 · 50.7 · n=663306-2306-2707-0107-0507-0907-1307-14

2 of 4 independent 14-day segments ready for trend judgment.

What explains the change

Experience change and discussion-mix change are separated. These are observational contributions, not proof of cause.

Largest negative experience contributions: General text, Coding, Reasoning.

Category Baseline → current Category change Weight share Experience contribution Discussion-mix contribution
General text 46.9 → 33.9n=873 → 3330 −13.0 36.7% → 41.9% −5.11 −1.04
Coding 81.6 → 66.5n=285 → 1000 −15.2 14.1% → 14.1% −2.14 −0.00
Reasoning 93.2 → 88.8n=689 → 1774 −4.4 34.9% → 26.4% −1.36 −2.57
Speed & latency 62.1 → 49.0n=84 → 452 −13.1 3.9% → 6.1% −0.66 −0.11
Safety & refusals 13.1 → 8.1n=133 → 613 −5.0 5.9% → 8.6% −0.36 −1.31
Roleplay / creative 81.5 → 76.7n=93 → 208 −4.9 4.1% → 2.7% −0.17 −0.27
Image / vision not enough datan=6 → 21
Local deploy not enough datan=3 → 3

Unresolved contribution from categories below the sample threshold: +0.10 pts.

Lifecycle scores use the same experience-signal weights as the main index, without sample shrinkage after n=30. They measure public user perception, not model capability or backend causes.