Sawfwair/MiniMax-H3-Ref2VA-MLX-8bit

MiniMax-H3 Ref2VA MLX 8-bit for mere.run

This is the self-contained, inference-ready MiniMax-H3 Ref2VA artifact for native Swift/MLX inference in mere.run. It generates synchronized 24 fps RGB video and 32 kHz stereo audio from an ordered presentation of image, video (including its soundtrack), and audio references.

The transformer is MLX affine INT8 with group size 64. Eight-bit is the lowest published Ref2VA precision because lower precision did not meet the visual quality bar. The Qwen3-VL conditioner is also MLX affine INT8/group-64; the video VAE is FP16 and the audio VAE is FP32.

The package includes a source-bound 31-point AdaLN cache. mere.run validates it against the immutable transformer SHA-256, uses the exact released schedule, and can omit the transformer's schedule-only AdaLN/time-embedding weights at load time. No post-pull model optimize step is required.

Install

mere.run model pull video-minimax-h3-ref2va-mlx --accept-model-license

The managed download is explicit and never occurs automatically. Review LICENSE before pulling or using the model.

Generate

Reference order is semantic and is preserved exactly:

mere.run video generate \
  "keep the subject, follow the camera move, and match the vocal rhythm" \
  --model video-minimax-h3-ref2va-mlx \
  --reference image:./subject.png \
  --reference video:./camera-and-soundtrack.mp4 \
  --reference audio:./voice.wav \
  --num-frames 124 \
  --output ./referenced-h3.mp4

The released limits are 12 references total: at most 9 images, 3 videos, and 3 audio clips. Audio cannot be the only reference type. Ref2VA regenerates synchronized audio conditioned by the presentation; it does not copy the input waveform into the output.

Provenance

The Ref2VA transformer was converted from exactly:

• repository: Comfy-Org/MiniMax-H3
• revision: fd70b39279d1ae6eb214c903f53e1bec3af19a77
• file: diffusion_models/minimax_h3_ref2va_int8_convrot.safetensors
• bytes: 34,038,894,550
• SHA-256: 9eef934046a0671bc8a5daf87100705e1478419c574cfde70c50fbe6885f76a9

The converter validates each tensor's embedded ConvRot metadata. The source uses group 256 for 200 transformer matrices and group 64 for 50 AdaLN matrices. It reverses each regular-Hadamard basis independently, then packs the restored weights as MLX affine INT8/group-64.

The resulting transformer is exactly:

• bytes: 36,024,412,656
• SHA-256: 234f22f69f8d40d6ed81cceed8259fa287f3c9417d40fba5274e3a7aa84e18a2

transformer.conversion.json, SOURCE_MANIFEST.json, MODIFICATIONS.md, and SHA256SUMS preserve the source, toolchain, transformation, and output evidence. The conditioner, VAEs, and tokenizer are exact files from Sawfwair/MiniMax-H3-FL2VA-MLX-4bit@e1244ad93d60c737c7e0f065a1c9372f3de7caf8.

Validation

The corrected artifact passed:

• source-operation parity before MLX requantization (relative L2 ~= 1.7e-6);
• packed transformer linear parity after MLX INT8 requantization (relative L2 ~= 0.00733);
• exact installed-cache parity against live three-modality AdaLN evaluation at the released schedule point used by the regression gate (zero maximum output error for both video and audio);
• a coherent 256x160/22-frame native Ref2VA smoke;
• a 512x320/124-frame native synchronized-A/V run on Apple M4 Max, producing the included railway-platform example and an intelligible generated dialog transcript.

The full validation run took 1,724.17 seconds with maximum acceleration. This establishes correctness and usable output, not speed parity with FL2VA or LTX.

License

The model weights are governed by the MiniMax H3 Community License Agreement, not the Apache-2.0 license of the mere.run source code. Preserve LICENSE, NOTICE, and MODIFICATIONS.md, and comply with all territorial, downstream, commercial, and acceptable-use requirements.