⚠️ REDOWNLOAD REQUIRED (2026-09-05): This bundle's MTP auxiliary files were reorganized. If you downloaded this model before 2026-09-05, please delete your local copy and re-download the entire repository — mixed old/new files can fail the runtime's bundle integrity check.
⚠️ Tool calling — parser notice. This model emits Qwen-XML tool calls (
<function=name><parameter=...>inside<tool_call>), not Hermes JSON. This bundle declarestool_parser: "qwen"in its config — use a runtime whose tool parser accepts the Qwen XML dialect. Hermes-only parsers will fail on tool turns (dropped/mangled calls). Runtime support for the correct parser is rolling out; until your runtime has it, expect tool-use errors.
JANGQ-AI/Qwen3.8-Flash-Next-JANG_4M
The recommended quality/size balance — median KL 0.0042 vs bf16 at 96.0 GiB (~73 GiB resident with the SSD-served table).
A JANG bundle of Qwen/Qwen3.8-Flash-Next — the Qwen4-architecture preview: a 125B mixture-of-experts (512 experts, 6B active) with a 51B hashed n-gram embedding, Gated DeltaNet + Qwen Sparse Attention hybrid layers, gated-residual streams, and vision+video towers — quantized for Apple Silicon / MLX. Text, image and video weights are all present in this exact bundle. Native multi-token-prediction head preserved (4-bit).
Best experienced in vMLX. This bundle's layout — the SSD-served n-gram table, per-module mixed precision, and the native MTP head — is designed for the vMLX serving path. Access is gated (manual approval) while runtime support rolls out.
Quality (measured, 5,931 held-out positions vs bf16)
| Tier | Size | RAM w/ SSD-table | median KL | top-1 | top-5 | top-10 |
|---|---|---|---|---|---|---|
| JANG_1L | 59.8 GiB | ~41 GiB | 0.0362 | 86.7% | 97.5% | 98.8% |
| JANG_2L | 65.3 GiB | ~48 GiB | 0.0260 | 88.2% | 98.2% | 99.0% |
| JANG_4S | 71.8 GiB | ~53 GiB | 0.0161 | 89.4% | 98.7% | 99.4% |
| JANG_4M | 96.0 GiB | ~73 GiB | 0.0042 | 94.4% | 99.7% | 99.9% |
| JANG_6S | 106.3 GiB | ~83 GiB | 0.0035 | 94.7% | 99.7% | 99.9% |
Margin-conditioned flip curves are monotone-decreasing on every tier — quantization noise lives in the reference model's own uncertainty band, with zero disagreement at high-confidence positions on the upper tiers.
The n-gram table & memory — SSD caching, fixed and fast
The 51B n-gram embedding streams directly from SSD on supporting runtimes (16 row-reads per token) — the "RAM w/ SSD-table" column above is the true resident footprint in that mode. Early runtime builds throttled in this mode; SSD-table caching is now fixed: decode runs at full speed with the table on disk — 40+ tok/s on an M5 Max for the 4-bit tier — so the biggest tiers fit comfortably on 64–128 GB machines without giving up the table.
What's in the bundle
- Vision + video: the full vision tower and both image and video preprocessors ship in this exact bundle — image-text-to-text and video understanding work out of the box on supporting runtimes (image and video token ids, mRoPE positions, and the merger are all present).
- Multi-token prediction: the model's native MTP head is preserved (trained multi-step). Enables self-speculative decode on supporting runtimes.
- Thinking + agentic: thinking mode on by default with three reasoning efforts and preserved thinking history; Hermes-style tool calling; the instruct preset gives direct non-thinking responses.
- Long context: 262,144 tokens native, extensible to 1M with YaRN.
Serving contract
- Thinking mode ON by default:
temperature=1.0, top_p=0.95, top_k=20 - Instruct mode:
temperature=0.7, top_p=0.80, top_k=20, presence_penalty=1.5 - Reasoning efforts
low / medium / xhigh(default xhigh) andpreserve_thinking(default on) via chat-template kwargs - Context 262,144 native, extensible to 1M with YaRN
- EOS
[248046, 248044]· tool calls: Hermes-style<tool_call>
Quantized and validated by Jinho Jang — eric@jangq.ai