MiniMax Opens H3-Image, Except to Americans
MiniMax released open-weight H3-Image, a 16-megapixel model it says renders in 10 seconds, but its license excludes the US, EU, UK and South Korea.
Read the storyMiniMax
Foundation-model company developing text, video, speech and music systems and AI-native applications.
Follow MiniMax
Explore the coverage Get AI Secret’s free daily briefing, including MiniMax and the wider AI industry.
At a glance
Foundation-model company developing text, video, speech and music systems and AI-native applications.
- Founded
- Early 2022
- Focus
- Text, video, speech and music
- Products
- MiniMax Code; Design; Audio; Hailuo; Talkie
- Reference
- MiniMax
Reviewed 19 October 2026. Dates retain the precision of the source; retrospective disclosures are identified in the event. Recognized revenue and annualized revenue are separate. Public reports use both ARR and run-rate terminology.
Company milestones
Revenue, funding, product launches and major company events, newest first. Expand the earlier milestones for more history.
H1 sales reach $116.6M
Revenue grows 283.1% year over year and exceeds the entire 2025 total. Enterprise and Open Platform services contribute $73.9M of this total.
Source: MiniMaxReported annualized revenue reaches $800M
MiniMax’s August annualized revenue reportedly grows from February’s $150M.
Source: The Information
Show earlier milestonesHide earlier milestones2026–2022 · 7 milestones
Earlier revenue, financing and company milestones.
Introduces MiniMax H3
The video model expands the company’s multimodal platform and open-weight deployment options.
Source: MiniMaxIntroduces MiniMax M3
The model advances coding, agentic workflows and professional-work applications.
Source: MiniMaxFY2025 sales rise to $79.0M
Annual recognized revenue grows 158.9% from $30.5M in 2024. This is distinct from the later ARR disclosures.
Source: MiniMax annual resultsDiscloses ARR above $150M in February
Management’s earnings-call figure reflects a forward annualized pace, rather than sales recognized during 2025.
Source: MiniMax earnings-call transcriptLists in Hong Kong after a $619M IPO
The offering transitions MiniMax from private funding into public equity markets.
Source: CB InsightsOpen-sources MiniMax-01
Text and visual-language models extend the company’s long-context and multimodal research platform.
Source: MiniMaxMiniMax is founded
The company begins developing multimodal foundation models and later commercializes them through consumer products and APIs.
Source: MiniMax
The latest on MiniMax
Newest firstRelated reporting
Briefs, articles and relevant sections, newest first. Explore older coverage below.
More coverage
1 min read · Members
MiniMax Credits Others for Shrinking M3.1 Flash
MiniMax thanked outside builders whose quantized versions, including a 143.9 GB 2-bit build, let its 428 billion parameter M3.1 Flash run locally.
7 min read · Members
Microsoft Offers Buyouts to 60,000 People
Plus: Anthropic's IPO Is an $800 Billion Lease, Meta Can't Wait for Its Own Stock
1 min read · Members
Beijing Clears DeepSeek to Keep Selling Abroad
Six days after China proposed AI export controls, a state media account said DeepSeek and MiniMax had registered, while API and app sales stay unrestricted.
1 min read · Members
China's AI Stocks Lose $480 Billion
China's AI stocks lost $480 billion in four months, MiniMax fell below its IPO price, and the STAR50 still trades near 100 times earnings.
8 min read · Members
Sam and Dario Sell Legroom
The ten smartest AI models carry only 6% of the traffic on an open platform.
2 min read · Members
MiniMax Unmasks Space Bunny as M3.1 Flash
MiniMax revealed the anonymous Space Bunny model as M3.1 Flash, a cheaper update it says cuts agent task costs 70%, with self-reported benchmark scores.
1 min read · Members
MiniMax H3 Launches Again, in Jiangxi
Five weeks after MiniMax open-sourced H3, a Jiangxi development zone staged its local launch, offering district companies 20% to 30% off.
1 min read · Members
MiniMax H3 Max Makes Live AI Video
MiniMax's H3 Max renders a five-second clip with audio in under three seconds, fast enough for live AI video streams.
7 min read · Members
Musk Thinks AGI Is Brute Force
Plus: A Corporate Guillotine, DeepSeek Finally Needs Money
