Modelpublic

Shiftedx/qwopus3.8-27b-flash-mxfp4-vision-mtplx

sourceHugging Faceapache-2.0updated 22d agoView on Hugging Face

6likes969downloads

Model Card

Qwopus3.8 27B Flash · MXFP4 Vision MTPLX

Apple-Silicon MLX release of Jackrong/Qwopus3.8-27B-Flash: an MXFP4 text body paired with the source model's exact BF16 vision tower and exact native BF16 MTP sidecar.

Artifact

Component	Format
Text body	MXFP4
Vision tower	BF16, 333 tensors
Native MTP sidecar	BF16, 15 tensors, depth 3
Source revision	`44d24e8cb20ceb3cdf4fe200b5a0afd970ee748a`
Qualified runtime	MTPLX 2.11.1 / MLX 0.32.2

The conversion used the pinned BF16 safetensors source—not a requantization of the GGUF release.

Run

bash

mtplx serve --model Shiftedx/qwopus3.8-27b-flash-mxfp4-vision-mtplx \
  --download --profile sustained --depth 3 --no-auth

This starts the local OpenAI-compatible endpoint. The packaged tokenizer, processor, vision weights, chat template, and MTP runtime contract are included.

Local qualification

One cold 192-token code row, seed 42, on an M4 Max 64 GB with automatic fan control:

AR	MTP D3	D3 / AR
26.43 tok/s	56.74 tok/s	2.15×

AR/D1/D2/D3 output checks passed. MTPLX inspection reported verified, runnable, and 15/15 MTP tensors. Strict text load returned the expected 42; vision identified the test shape/color exactly. These are narrow local qualification results, not a cross-hardware benchmark.

Integrity and notes

—Vision SHA-256: 7a252e5c7bb02ed6533f905a566ab037356ffb4949d04df8d6d2a3fa5e3af6b9
—MTP SHA-256: 4468f39621de68a19ffd0bcb2e2e2f352205def7436a625b3427e3752866c287
—SHA256SUMS covers every packaged artifact file except itself.
—The upstream tokenizer's Mistral-regex warning is preserved for exact source lineage. The packaged chat template expects string-valued message content.
—Quantization can change quality. Apply the source model's license and limitations to this derivative.