Claude Sonnet 5

NewClaude 5

Ranks #8 of 35 on reading real contracts into structured billing data, -0.6 pts vs GPT-5.5.

75.4%

Accuracy

#8 of 35

6.2%

Hallucination

HIGH-confidence & wrong

$200

Cost / 1k contracts

$0.20 each · via OpenRouter

106.7s

Median latency

p90 154.2s

±0

Run-to-run σ

1 runs

0.38

Value

acc. pts per $/1k

Tokens per contract

Input36,515

Output12,714

Reasoning0

Reliability 100% · valid structured output across 30 calls.

Want the full picture? How we score