Nova 2 Lite

NewNova

Ranks #33 of 43 on reading real contracts into structured billing data — -18.7 pts vs GPT-5.5.

50.6%

Accuracy

#33 of 43

36%

Hallucination

HIGH-confidence & wrong

$11.2

Cost / 1k contracts

$0.01 each · via OpenRouter

13.3s

Median latency

p90 17.1s

±9.6

Run-to-run σ

3 runs

4.5

Value

acc. pts per $/1k

Tokens per contract

Input20,974

Output1,981

Reasoning0

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →