MiniMax M3

NewMiniMax

Ranks #37 of 43 on reading real contracts into structured billing data — -25.2 pts vs GPT-5.5.

44.1%

Accuracy

#37 of 43

33.6%

Hallucination

HIGH-confidence & wrong

$10.7

Cost / 1k contracts

$0.01 each · via OpenRouter

55.5s

Median latency

p90 390.1s

±11.8

Run-to-run σ

3 runs

4.14

Value

acc. pts per $/1k

Tokens per contract

Input21,125

Output3,594

Reasoning0

Reliability 100% · valid structured output across 18 calls.

Want the full picture? How we score →