Claude Fable 5.2 on the quick suite
A typical tickets-test run cost an estimated $1.593. All 21 of its runs passed.
Notes
Typical means the middle run, or the midpoint of the two middle runs.
Time to finish
One short explain task, timed six times. The median passed run took 11.9 s. Lower is faster.
Every run, one row each
| Model | Tool | Run 1 | Run 2 | Run 3 | Run 4 | Run 5 | Run 6 |
|---|---|---|---|---|---|---|---|
| Claude Fable 5.2 | Claude Code | 11.2 s | 14.2 s | 9.4 s | 12.2 s | 13.0 s | 11.6 s |
The table scrolls sideways on a small screen.
Source: WorkloadBench at workloadbench.brenhq.com, updated 9 November 2026. 6 runs, medium effort, in Claude Code 2.1.321.
Notes
How time is measured: Method.
Watch the race
Claude Fable 5.2, six timed runs
Every time-to-finish run, replayed together.
- Claude Fable 5.2: 9.4 to 14.2 s
Source: WorkloadBench at workloadbench.brenhq.com. 6 runs in 1 lanes, played at 3x speed. The clock counts real seconds. I timed each run on its own. Here they start together, and each dot moves at a steady speed to its real finish time.
Real app work
A bug hunt and a tickets test, three runs each. Every run passed with full marks.
Bug hunt
Tickets
Source: WorkloadBench at workloadbench.brenhq.com, updated 9 November 2026. 6 runs, medium effort, in Claude Code 2.1.321.
Explain task
All three runs passed. Each run's cost at list prices when run.
Source: WorkloadBench at workloadbench.brenhq.com, updated 9 November 2026. 3 runs, medium effort, in Claude Code 2.1.321.
Output speed
A walkthrough of at least 600 words. The median run was 126 tokens a second. Higher is faster.
Source: WorkloadBench at workloadbench.brenhq.com, updated 9 November 2026. 6 runs, medium effort, in Claude Code 2.1.321.
Notes
A run counts when it passed and its answer streamed in 20 or more pieces, and a median needs three runs that count.
How output speed is timed: Method.
Every check, price and setting
Anthropic released Claude Fable 5.2 on 9 November. Claude Fable 5.2 ran in Claude Code 2.1.321 at medium effort.
Hidden tests are tests the model never sees. A run passes its gate when every test passes, its own new ones included, and no existing test file was changed, renamed or deleted. New test files are fine.
| Check | Claude Fable 5.2 |
|---|---|
| Explain task: runs passed | 3 of 3 |
| Bug hunt: bugs fixed of 20, each run | 20, 20, 20 of 20 |
| Bug hunt: runs that failed the gate | 0 of 3 |
| Bug hunt: median cost of a passed run | $1.185 |
| Tickets: hidden tests passed, each run | 17, 17, 17 of 17 |
| Tickets: runs that failed the gate | 0 of 3 |
| Tickets: median cost of a passed run | $1.593 |
| Time to finish, explain task: median passed run | 11.9 s |
| Output speed, a walkthrough of 600 words or more: median o200k tokens a second | 126 |
The table scrolls sideways on a small screen.
Costs are estimates at list prices when run, from each passed run's tokens. Method lists the rates.