MiniMax
9 published articles
Artificial Intelligence
The missing 'ums' and 'uhs' that finally make AI speech sound human
MiniMax's Speech 2.8 brings native filler words, breathing, and hesitation to AI voice synthesis, solving the 'too perfect' problem that has long made synthetic speech feel robotic. The model also delivers 10-second voice cloning and cross-language accuracy improvements.
2026-07-11
AI Labs & Research
MiniMax's M3 just wrote its own CUDA kernel, and opened the code
MiniMax M3 scores 83.5 on BrowseComp, edges past Opus 4.7, and handles up to 1M tokens natively. In a remarkable autonomy test, it self-optimized a GPU kernel from 7.6% to 71.3% peak utilization without human intervention.
2026-07-09
AI Model Release
MiniMax launches M2.7 model with strong software engineering and office productivity skills
MiniMax's M2.7 model delivers strong results in software engineering benchmarks and professional office tasks, with a 97% skill adherence rate on complex instructions and an ELO score of 1495 on GDPval-AA.
2026-07-08
AI Speech
Minimax speech 2.8 brings human warmth to AI voices with native filler words and high-fidelity cloning
MiniMax Speech 2.8 adds natural filler words, 10-second voice cloning, and studio-quality audio to close the gap between synthetic and human speech. The model fixes the 'too perfect' problem that makes AI voices sound robotic.
2026-07-08
AI Video Generation
MiniMax launches Hailuo 2.3 video model with enhanced motion and style support
MiniMax's Hailuo 2.3 video model improves motion rendering, expands style support for anime and CG, and delivers more natural facial expressions. The companion Media Agent lets users create videos from text, images, and audio with a single click.
2026-07-07
Artificial Intelligence
MiniMax's product blitz: new models for code, music, and video in one broad refresh
Chinese AI startup MiniMax launches MiniMax M3, Hailuo 2.3, MiniMax Code, and new speech/music models, broadening its product lineup in a competitive landscape.
2026-07-06
Artificial Intelligence
MiniMax M3 broke the open-weight ceiling with a 9.4x CUDA speedup and zero human help
MiniMax M3 delivers a 9.4x CUDA kernel speedup, beats Opus 4.7 on BrowseComp, and autonomously replicated an ICLR paper. All in an open-weight package.
2026-07-06
Artificial Intelligence
MiniMax's new M2.5 coding model tops the benchmark at 5% of the price
MiniMax's M2.5 model tops the Multi-SWE-Bench coding benchmark, beats mainstream models on workspace tasks, and costs a tenth to a twentieth of competitors. Open-source weights are on HuggingFace.
2026-07-04
AI Research & Development
Ma Jiaqi taught MiniMax engineers a hard lesson about forgotten tokens
MiniMax's internal investigation into why its M2 model couldn't output the name 'Ma Jiaqi' revealed a structural mismatch between pre-training vocabulary and post-training data distribution. The root cause: low-frequency tokens' lm_head vectors drift during SFT, losing generation ability while retaining understanding. A full-vocabulary coverage fix resolved the issue and also mitigated language mixing in Japanese.
2026-07-04