Mistral Large

NewMistral

Ranks #14 of 35 on reading real contracts into structured billing data, -5 pts vs GPT-5.5.

71%

Accuracy

#14 of 35

19%

Hallucination

HIGH-confidence & wrong

$18.7

Cost / 1k contracts

$0.02 each · via OpenRouter

43.3s

Median latency

p90 49.1s

±0

Run-to-run σ

1 runs

3.79

Value

acc. pts per $/1k

Tokens per contract

Input24,977

Output4,171

Reasoning0

Reliability 100% · valid structured output across 30 calls.

Want the full picture? How we score