Qwen3.7 Plus

Qwen3

Ranks #18 of 43 on reading real contracts into structured billing data — -5.5 pts vs GPT-5.5.

63.8%

Accuracy

#18 of 43

22.8%

Hallucination

HIGH-confidence & wrong

$20.3

Cost / 1k contracts

$0.02 each · via OpenRouter

234.9s

Median latency

p90 1274.2s

±10.3

Run-to-run σ

3 runs

3.15

Value

acc. pts per $/1k

Tokens per contract

Input23,383

Output9,986

Reasoning5,694

Reliability 83.3% · valid structured output across 18 calls.

Want the full picture? How we score →