MiniMax Group (0100.HK) NDR highlights & positive read-across from DeepSeek V4 peak-hour price hike; Buy

Research evidence excerpt

MiniMax Group (0100.HK) NDR highlights & positive read-across from DeepSeek V4 peak-hour price hike; Buy

s seeking fresh graduate talents (from top China

and overseas universities), and an increasing number of senior experts from vertical

domains through its 10X talent plan, who can translate their industry know-how into

model training and real-world task optimization.

Our take on DeepSeek’s upcoming peak-hour price hike from mid-July: Positive

read-through for MiniMax M3’s performance/price competitiveness

DeepSeek announced this week that its V4 o %cial version is set for launch in mid-July,

alongside the introduction of peak/o ’-peak API pricing to better allocate resources and

improve service stability. V4 Pro/Flash non-peak pricing remains unchanged, while peak

hour (9am-12pm/2pm-6pm China time) will be charged at 2X non-peak rates, implying

blended pricing of US$0.35/US$0.12 per 1M tokens due to strong Chinese AI model

demand that is increasingly causing signi Hcant compute tightness in work/productivity

scenarios. We view it as an early sign that China AI model companies’ aggressive pricing

since late-April 2026 (at or below zero gross margins for some players) may be moving

toward a more rational phase into July in re ection of inference costs. We see MiniMax

M3’s blended pricing at US$0.22 per 1M tokens as attractive vs. performance, at

signi Hcantly above peers’ gross margins helped by its higher proportion of optimized

self-built computing and cost-e %cient architecture (at smaller activated parameters)

that can support more sustainable unit economics for MiniMax’s M3 model family.

We continue to see MiniMax as well positioned amongst Chinese AI model companies to

capture the signi Hcant global TAM growth potential across text/coding, multi-modal