Claude Fable 5.1

NewClaude Fable 5.1

Ranks #1 of 35 on reading real contracts into structured billing data, +3.4 pts vs GPT-5.5.

79.4%

Accuracy

#1 of 35

2.6%

Hallucination

HIGH-confidence & wrong

$674

Cost / 1k contracts

$0.67 each · via OpenRouter

61.2s

Median latency

p90 73.8s

±0

Run-to-run σ

1 runs

0.12

Value

acc. pts per $/1k

Tokens per contract

Input36,517

Output6,167

Reasoning946

Reliability 100% · valid structured output across 30 calls.

Want the full picture? How we score