MiniMax-H3-Qwen3-VL-NVFP4
An NVFP4 (4-bit) Qwen3-VL-32B text encoder for MiniMax-H3, made with
convert-to-quant. It is a drop-in replacement
for the H3 conditioning text encoder (ComfyUI CLIPLoader, type minimax). The H3 video weights are
unchanged — only the text encoder is quantized to 4-bit, to bring the full H3 stack onto consumer GPUs.
Quantization
- Quantized with
ctq --nvfp4 --qwen_vlm --comfy_quant(weight-only PTQ, comfy_quant layout). - Protected (kept bf16): token embeddings, first & last language layers, and the full vision
tower (
--qwen_vlmfilter). The middle language-layer linears go to NVFP4 (FP4 E2M1, block 16). - Size: 16.64 GiB / 17.86 GB (bf16 source is 47.97 GiB) — fits a 24 GB consumer GPU with ComfyUI offload.
What's different
This encoder deliberately keeps more of the network in bf16 than a minimum-size 4-bit encoder does.
The --qwen_vlm filter protects — at bf16 — the token embeddings, the first and last language
layers, and the entire vision tower, quantizing only the middle language-layer linears to NVFP4.
The result is 16.64 GiB, about 2 GiB above the most aggressively quantized community NVFP4
encoders (≈14.6 GiB, e.g. the AWQ-quantized encoder in Comfy-Org/MiniMax-H3), in exchange for
leaving the precision-sensitive layers (embeddings, boundary layers, vision) untouched. It is a
fidelity-first tradeoff: spend ~2 GiB more VRAM to keep the layers that most affect conditioning at
full precision. Everything still fits a 24 GB card with offload.
Quantization fidelity (4-bit vs bf16)
Everything is held constant except encoder precision — the bf16 encoder vs this NVFP4 encoder, same prompts, same seeds, same DiT — over a 16-clip paired set.
Verdict — perceptually and temporally equivalent to bf16:
| Metric | Value | What it measures |
|---|---|---|
| LPIPS ↓ | 0.099 | Learned perceptual per-frame distance |
| SSIM ↑ | 0.891 | Structural per-frame similarity |
| Temporal flicker (mean inter-frame diff) | 13.75 → 14.12 (+2.7%) | Frame-to-frame stability across the clip set |
LPIPS ≈ 0.10 with SSIM ≈ 0.89 means the 4-bit encoder's frames are perceptually the same as bf16's, and the near-identical flicker (13.75 vs 14.12) shows temporal behavior is preserved — the 4-bit conditioning adds no visible shimmer or instability. These are the fidelity conclusion.
FVD — internal relative reference only, not a fidelity verdict. i3d FVD over the same 16 paired clips is 1059.5. This is not a published-scale FVD number: FVD estimates a high-dimensional feature covariance and needs large samples (papers typically use 2048+); at n = 16 the covariance is under-determined, which inflates the absolute value and makes it non-comparable to reported FVD. A meaningful reading would require a same-precision noise floor (bf16 vs bf16 at a different seed); we report the raw figure only for transparency and rely on the perceptual and temporal metrics above.
PSNR over the paired clip is 24.85 dB — low only because a tiny conditioning change makes the diffusion sampler diverge pixel-wise while staying perceptually equivalent, which is exactly why LPIPS/SSIM (not PSNR) are the right lens here.
Verified
| Item | Result |
|---|---|
Loads in ComfyUI (CLIPLoader type minimax) | Yes |
| Generation quality after 4-bit quantization | Clean 720p; perceptually equivalent to bf16 (SSIM 0.891, LPIPS 0.099) |
| Character identity-lock (ref2va, 3-view) | Consistent |
| Peak VRAM, full 4-bit stack, 24 GB budget | ~25.7 GiB (runs on RTX 4090 with offload) |
| Speed | ~29 s/step at 24 GB; ~2 min/clip with the 4-step turbo LoRA; faster on 32 GB+ |
NVFP4 runs fastest on Blackwell; on Ada (RTX 4090) a cu130 torch build enables the
comfy_kitchenFP4 kernels (otherwise emulated and ~2.3× slower).
Usage (ComfyUI)
- Put
qwen3vl_32b_minimax_h3_nvfp4.safetensorsinComfyUI/models/text_encoders/. - Download the companion 4-bit weights:
- DiT:
minimax_h3_ref2va_pruned_nvfp4.safetensors→diffusion_models/(lilcheaty/MiniMax-H3-NVFP4) - VAE:
minimax_h3_video_vae_fp16+minimax_h3_audio_vae_fp32→vae/(Comfy-Org/MiniMax-H3) - Turbo (optional, 4-step):
minimax_h3_fl2v_turbo_4step_v0.1.safetensors→loras/(Gluttony10/MiniMax-H3-INT8-CONVROT)
- DiT:
- In the H3 workflow set
CLIPLoaderto this encoder andUNETLoaderto the NVFP4 DiT.
Notes
Weight-only 4-bit quantization of the text encoder only; the encoder's behavior is otherwise that of the base Qwen3-VL-32B / MiniMax-H3 release. Follow the MiniMax Community License and local law.
Credits & license
- Base video model: MiniMax-H3 (MiniMax Community License); encoder base Qwen3-VL-32B; bf16 encoder packaging: Comfy-Org/MiniMax-H3.
- Quantizer: convert-to-quant (silveroxides). Companion 4-bit DiT: lilcheaty; turbo LoRA: Gluttony10.
- License inherited from the base model.