Gemini 3.5 Flash

Gemini 3

Ranks #2 of 43 on reading real contracts into structured billing data — +9.3 pts vs GPT-5.5.

78.6%

Accuracy

#2 of 43

13.8%

Hallucination

HIGH-confidence & wrong

$132

Cost / 1k contracts

$0.13 each · via OpenRouter

49.9s

Median latency

p90 58.4s

±2.9

Run-to-run σ

3 runs

0.6

Value

acc. pts per $/1k

Tokens per contract

Input24,911

Output10,524

Reasoning6,555

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →