Comfy-Org republishes MiniMax-H3-Image weights for ComfyUI
The pack adds int8 and NVFP4 files plus Qwen encoders, under a license that restricts high-revenue hosted use.
Stories per day · 30 days
2 RSS
Relatedopen-weights7video-generation7huggingface5lora4comfyui3mixture-of-experts3benchmarks2image-generation2quantization2agentic-models1agent-skills1audio1civitai1deepseek1diffusion1distributional-rl1empirical-bayes1entity-consistency1
37 stories, newest first · one card per story
Stories tagged #minimax
Video Rebirth publishes HyperFlow H3 Flash, a compressed MiniMax-H3
The 13.7B–13.9B model makes 5.2–15 second video with stereo audio in 4–8 steps, under a restricted MiniMax license.
MiniMax releases open-weight MiniMax-H3-Image for generation and editing
The checkpoint outputs about 16-megapixel images in 6 steps and accepts up to 16 reference images.
MiniMax open-sources MCode Stack of 15 agent skills
The supplied page text is a cookie notice and does not describe the skills or license.
Open-H3-Regenerate-2K upscales MiniMax-H3 drafts to 2560×1440
Authors report a five-second clip falling from 87.26 seconds to 22.45 seconds on one B200 with the fast decoder.
3
PRISM merge of Minimax-H3 Singularity and 10Eros-Max trends on Hugging Face
The fp16 file merges two community Minimax-H3 finetunes without mergekit, and the card adds later int8 tests.
7
Paper gives exact global minimax risk near nonidentification
Conditions let one estimator match the leading constant uniformly; two models fix fifth- and sixth-order risk constants.
Licon VBVR LoRA adapts MiniMax-H3 for image-to-video
The adapter, H3-VBVR-I2V.safetensors, is trained on rebuilt VBVR data and loads in a MiniMax-H3 ComfyUI workflow.
AnyGameRenderer restyles game footage while keeping entities
A LoRA on 33B MiniMax-H3 raises temporal entity match from 50.9% to 62.9% on 59 held-out cases.
NPMLE shown minimax optimal for univariate Gaussian mixtures
With compact support, a constrained NPMLE reaches the slightly faster rate log(n)/(n log log n).
Paper proves near-minimax rates for quantile distributional RL
A quantile fixed-point estimator matches local asymptotic minimax risk, with resolution-independent rates when reward densities are bounded away from zero.
DeepSeek V4.1 Flash 1017 leads open-weight value rankings
Artificial Analysis scores the 552B MoE at 45, up six points, tying GLM-5.3 among open weights.
1
FastVideo publishes ComfyUI weights for FastH3 Omni
BF16 and INT8 ConvRot checkpoints use eight steps and need MiniMax-H3 text, video and audio VAEs.
SolarRenderer adapts MiniMax-H3 for four-step structure-conditioned video
Two LoRAs keep a frozen 33B backbone and render about 12.3 times faster than 50-step sampling.
MiniMax releases M3.1 Flash, with open weights due October 23
API pricing is $0.15 per million input tokens and $0.60 per million output tokens, with cached input at $0.015.
1
MiniMax launches M3.1 Flash, a 428B MoE agentic model
The listing prices input at $0.15 per million tokens; the fetched page body is only a cookie notice.
1
Stateful Harness splits rule execution from video world-model rendering
On 50 counterfactual twin pairs, MiniMax-H3 rises from 32% to 92% rule compliance with the harness.
MiniMax releases MiniMax-M3.1-Flash, a 428B multimodal MoE
About 23B parameters activate per token; inputs include text, image, and video, with a 1 million token context window.
2
Minimax-optimal Scaling Laws in Ridge Regression via Transfer Learning
Transfer from ridge-regression pretraining can achieve minimax-optimal scaling under power-law covariance shift.
WorldGuide: Goal-Directed Video World Model for Procedural Task Execution
WorldGuide couples a planner and executor for closed-loop, goal-directed procedural video generation.
Detecting a Shift Is Not Enough: Exact Minimax Limits of Linear Representation Repair
Exact minimax limits show linear shift repair needs far more calibration signal than detection.
MC-Sparse: Deconstructing and Closing the Dense-Sparse Attention Gap in Diffusion Transformers
Training-free MC-Sparse speeds diffusion-transformer denoising up to 2.32× with negligible quality loss.
MC-Sparse: Deconstructing and Closing the Dense-Sparse Attention Gap in Diffusion Transformers
MC-Sparse narrows the dense-sparse attention gap, speeding diffusion denoising up to 2.32× with little quality loss.
DMAD: Distribution Matching as Adversarial Distillation for Fast Visual Generation
DMAD distills few-step image and video generators via adversarial distribution matching, reporting top compared scores.