GPT-5.6 Luna

NewGPT-5.6

Ranks #12 of 43 on reading real contracts into structured billing data — -3.1 pts vs GPT-5.5.

66.2%

Accuracy

#12 of 43

17%

Hallucination

HIGH-confidence & wrong

$4.30

Cost / 1k contracts

$0.00 each · via OpenRouter

26.9s

Median latency

p90 147.5s

±1.8

Run-to-run σ

3 runs

15.28

Value

acc. pts per $/1k

Tokens per contract

Input21,576

Output3,627

Reasoning987

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →