 | |
| Company type | Private |
|---|---|
| Industry | Information technology |
| Founded | December 2021; 3 years ago (2021-12) |
| Founders |
|
| Headquarters | Shanghai,China |
Key people |
|
| Products |
|
Number of employees | 200+ (2025) |
| Website | minimax.io |
MiniMax is anartificial intelligence (AI) company based inShanghai, China. As of 2024, it has been dubbed one of China's "AI Tiger" companies by investors.
MiniMax was founded in December 2021 by severalcomputer vision veterans fromSenseTime. When it first started out, it received funding fromMiHoYo.[1][2][3][4]
In March 2024,Alibaba Group led a $600 million financing round for MiniMax giving it a valuation of $2.5 billion. Other investors of MiniMax includeHillhouse Investment,HongShan,IDG Capital andTencent.[1][5]
MiniMax's first product was Glow which was launched in October 2022. The app allowed users to create virtual characters, give them background stories and then chat with them about various topics. Only four months after launch, the app had over 5 million users. Due to filing issues, Glow was terminated in March 2023.[3][4]
Glow was relaunched under two new brands: Talkie, launched for international markets in June 2023, and Xing Ye (星野), introduced in September 2023 for the Chinese market.[4][5]
Talkie was profiledThe Wall Street Journal in July 2024. For June 2024 Talkie ranked fifth among the most-downloaded free entertainment apps in the U.S. More than half of Talkie's 11 million monthly active users were in the U.S. Other popular markets included Philippines, the U.K. and Canada. Talkie offered AI-generated conversations with people such asDonald Trump,Taylor Swift,Elon Musk andLeBron James.[6]
A hoodie man walking forward in the rainy cyberpunk street, carrying an umbrellaIn March 2024, MiniMax launched Hailuo AI, a multimodallarge language model consumer platform that provides AI text and music-generating features.[2][4]
In September 2024, MiniMax launched video-01, atext-to-video model under Hailuo AI.[2] A review by Tom's Guide stated it roughly equivalent to Luma LabsDream Machine but not as good asRunway Gen-3.[7]
In December 2024,Broadcast reported that Hailuo AI can reproduce the logos of British television channelsChannel 4,Channel 5, andITV in its AI-generated videos.[8]
On 20 January 2025, MiniMax launched audio functions forHailuo AI.[9]
On 28 January 2025, MiniMax release the Model T2V-01-Director and I2V-01-Director, is the latest addition to the Hailuo AI Video-01 Model family, are designed to revolutionize video creation by putting professional filmmaking tools at fingertips.[10] T2V-01-Director and I2V-01-Director introduce a significant reduction in movement randomness, ensuring that every frame adheres strictly to creative intent. This heightened level of control allows for more accurate and nuanced storytelling.[11]
On 17 April 2024, MiniMax officially launched the ABAB 6.5 series, amixture of experts language model.[4]
In January 2025, Minimax unveiled the MiniMax-01large language model (LLM) family, which includes a general-purpose foundational model, the MiniMax-Text-01, and the multimodal MiniMax-VL-01 with visual capabilities.[12]
In April 2025, Minimax released Speech-02, a newtext-to-speech model. It supports over 30 languages and can process up to 200,000 characters at once.[13]
In June 2025, Minimax released MiniMax-M1, "the world's first open-weight, large-scale hybrid-attention reasoning model". It supports a context length of 1 million tokens, and the lightning attention mechanism in MiniMax-M1 enables efficient scaling of test-time compute – for example, compared to DeepSeek R1, M1 consumes 25% of the FLOPs at a generation length of 100K tokens.[14]
