LaunchDevelopers6 sources
AI Lab
MiniMax News
Official MiniMax announcements — model releases, product launches and research, each one summarized with every source covering it, by AIBriefs. RSS
EventBusiness1 source
MiniMax hosts generative AI and entertainment event in Tokyo
AnalysisVisual AI5 sources
Minimax H3 Singularity fine-tune released on HuggingFace
Community fine-tune of Minimax H3 adds HDR image quality, blur reduction, distant face restoration and VFX effects while retaining full base capability. The HuggingFace upload has drawn 72 likes.
EventAI Models8 sources
Saudi Arabia's Humain unveils AI model based on China's MiniMax
Humain built HUMAIN-M3 on MiniMax's M3 foundation, trained on over 1 trillion Arabic tokens. It aims to bring frontier-level capabilities to Arabic and regional dialects.
LaunchVisual AI1 source
H3 video generation runs locally on consumer hardware
AnalysisVisual AI1 source
Real-time interactive video generation built on vLLM-Omni and FastH3
AnalysisAI Models3 sources
MiniMax H3 open weights become most-downloaded MiniMax model
MiniMax says H3 open weights dropped one month ago; a Reddit user reports it became the most downloaded MiniMax model, and the most downloaded on ComfyUI, in about two weeks.
LaunchAI Models1 source
MiniMax H3 and H3 Max models launch on Vercel
LaunchAI Models2 sources
MiniMax H3 Max model released as open-weight
LaunchVisual AI2 sources
MiniMax Fast H3 v1 open-sourced, 14x faster video generation
MiniMax Fast H3 v1 generates 15s 768p video in 13s, 14x faster on a single GPU. Built with Hao AI Lab, Nua Lab, and NVIDIA FastGen teams.
LaunchAI Models4 sources
fal builds MiniMax H3-Max video model with co-designed inference
fal's H3-Max combines post-training with a co-designed inference stack to improve prompt adherence, visual quality, and speed. fal will release the weights of H3 Max.
EventDevelopers1 source
Ray Summit 2026 kicks off in San Francisco with AI builders
LaunchAI Models5 sources
FireRedTeam releases FireRedAudio and FireRedTTS3
FireRedAudio is a general-purpose audio language model using decoupled continuous representations for understanding and generation. FireRedTTS3 unifies speech generation and editing with semantically enriched speech representations.
EventDevelopers1 source
MiniMaxthon hackathon launches with three tracks over 14 days
LaunchAI Models1 source
MiniMax-M3: $0.018 per task, cheapest on agent benchmark
How-ToDevelopers1 source
Awesome MiniMax H3 Integrations index tracks ecosystem growth
AnalysisAI Models1 source
NVIDIA Sol Engine cuts MiniMax H3 latency to 14.93s
LaunchAI Models1 source
MiniMax M3 model gains support for SambaNovaAI hardware
EventDevelopers1 source
MiniMax and ComfyUI launch H3 challenge
LaunchDevelopers6 sources
TrueFoundry open sources TrueForge agent harness
TrueForge, an open source (MIT) AI agent harness from TrueFoundry, claims 30%-75% cheaper task completion than Claude Managed Agents. It launched on GitHub and already has nearly 4K stars, with partners including MiniMax and Together AI.
EventVisual AI1 source
AI character powered by MiniMax hits 1M views in a week
AnalysisAI Models1 source
Streaming video LoRA adapter for MiniMax H3 model proposed
LaunchAI Models2 sources
MiniMax H3 model becomes unlimited on Runway
LaunchDevelopers1 source
SGLang/RadixArk launches Miles v0.1
LaunchAI Models1 source
MiniMax and Vercel demonstrate AI-integrated workout tools
How-ToAI Models1 source
MiniMax H3 video generation model workflows shared for RTX 4090
LaunchVisual AI2 sources
MiniMax H3 generates consistent game sprite animations
MiniMax's H3 video generation model produces consistent walking motion for game sprites, a task where some image models struggle. The model is available for sprite animation use.