Claude Fable 5
17 of 17 task types · 14,547 records · 12,043 correct
80.1%
Suite Mean
unweighted avg of the headline tasks
legacy 11-task 84.8%
Evaluation Configuration
Reasoning
Adaptive — Adaptive thinking is always on and cannot be disabled. Effort set to max on 13 of the 17 tasks; on Citation Lookup, Citation Holding, Limitation Periods and Issue Spotting no effort setting is sent (provider default).
Max output tokens32,768
TemperatureAPI default
NotesTemperature is not sent (model rejects non-default values). API default applies.
Task Performance
| Task | Accuracy | |
|---|---|---|
| Cite Verify | 97.8% | |
| Statute QA | 81.0% strict · lenient 93.4% | |
| Jurisdiction Router | 89.5% | |
| Bijural Equivalence | 54.9% strict · lenient 78.9% | |
| Case Treatment | 94.8% | |
| Case Outcome | 89.4%† | |
| Standard of Review | 98.9% | |
| Doctrine Shift | 80.0% | |
| Regulation QA | 62.9%†‡ strict · lenient 62.9% | |
| Tribunal QA | 88.3% strict · lenient 88.3% | |
| Multi-Doc Reasoning | 80.0%† | |
| Temporal Reasoning | 91.7% | |
| Citation Lookup | 93.9% | |
| Citation Holding | 56.0% strict · lenient 56.0% | |
| Limitation Periods | 100.0% | |
| Issue Spotting | 58.7% every key issue · exact set 0.7% · lenient 55.3% | |
| Citation Resolution | 44.4%† |
† the model's shared thinking and answer token budget ran out before it wrote any answer.‡ the provider's safety filter blocked the answer, or the answer came back empty with no recorded finish reason.These answers stay in the denominator and are scored 0; the count is disclosed per cell. Empty-output policy: an answer that came back empty is never removed from its cell; it counts as not correct against the frozen record count, and the number of such answers and their cause are published with the cell.
- Case Outcome: † 3 answers hit the token budget before any answer token; scored 0
- Regulation QA: † 1 answer hit the token budget before any answer token; scored 0 · ‡ 1 answer blocked by the provider's safety filter (empty output); scored 0
- Multi-Doc Reasoning: † 1 answer hit the token budget before any answer token; scored 0
- Citation Resolution: † 202 answers hit the token budget before any answer token; scored 0