GPT-5.6 Sol

NewGPT-5.6

Ranks #9 of 43 on reading real contracts into structured billing data — -1.3 pts vs GPT-5.5.

68%

Accuracy

#9 of 43

8.7%

Hallucination

HIGH-confidence & wrong

$242

Cost / 1k contracts

$0.24 each · via OpenRouter

78.7s

Median latency

p90 119.5s

±1.7

Run-to-run σ

3 runs

0.28

Value

acc. pts per $/1k

Tokens per contract

Input21,564

Output4,460

Reasoning1,663

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →