Qwen3.7 Flash

NewQwen3

Ranks #34 of 43 on reading real contracts into structured billing data — -19.3 pts vs GPT-5.5.

50%

Accuracy

#34 of 43

38%

Hallucination

HIGH-confidence & wrong

$1.80

Cost / 1k contracts

$0.00 each · via OpenRouter

118.5s

Median latency

p90 130.3s

±0

Run-to-run σ

3 runs

27.82

Value

acc. pts per $/1k

Tokens per contract

Input21,404

Output8,885

Reasoning6,313

Reliability 83.3% · valid structured output across 6 calls.

Want the full picture? How we score →