Llama 4 Maverick

NewLlama 4

Ranks #28 of 43 on reading real contracts into structured billing data — -11.3 pts vs GPT-5.5.

58%

Accuracy

#28 of 43

27%

Hallucination

HIGH-confidence & wrong

$5.50

Cost / 1k contracts

$0.01 each · via OpenRouter

39.7s

Median latency

p90 1205.8s

±5

Run-to-run σ

3 runs

10.53

Value

acc. pts per $/1k

Tokens per contract

Input20,678

Output1,969

Reasoning0

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →