Unverified50% confidenceFactExact time
Mini-AGI is an open-source project positioned as training a model with dynamic continual learning capabilities from scratch on a consumer-grade GPU with only 8GB of VRAM.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/21/2026
First Seen
Valid until: 12/20/2026
Sources
Related Entities
Related Claims
UnverifiedOn a single consumer-grade GPU (such as an NVIDIA 3090), MiniMind can complete the full training process from random initialization to a conversational model in as little as 2 hours.67% similarUnverifiedThere is a native desktop agent built on GPUI that is in an early stage.66% similarUnverifiedEven graphics cards with around 8GB of VRAM can run the MiniMax open-source accelerated model.65% similarUnverifiedMulti-GPU training supports automatic device allocation, enabling larger models to be split across multiple graphics cards.65% similarUnverifiedThe team has opened an OpenAI-compatible free research-use API with support from Nvidia GPUs, rate-limited to 5 RPM64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/938206API
curl https://kongchang.com/api/v1/knowledge/claims/938206MCP
get_claim(id=938206)