MoE-aware, mixed-precision **APEX** quantization of Kwaipilot/KAT-Coder-V2.5-Dev, an agentic-coding fine-tune on the Qwen3.6-35B-A3B backbone (`qwen3_5_moe`: 40 layers, 256 routed + 1 shared expert, hybrid GatedDeltaNet linear-attention with periodic full attention). Kwaipilot reports SWE-bench Verified 69.40 for this size class, against Qwen3.6-35B-A3B's 64.40. **What makes this build different: it has a working MTP head.** KAT-Coder ships `mtp_num_hidden_layers: 0` — no multi-token-prediction head at all, so no speculative decoding is possible out of the box, and that is true of the vendor release and of every other quant of this model I am aware of. This repo transplants Qwen3.6-35B-A3B's own trained MTP head onto KAT-Coder and quantizes the result. Measured below: **2.03x on a hard agentic-coding suite, with correctness unchanged.** Full draft-depth x temperature surface, measured on `-v2`. One long-context prompt (~82 KB), `max_tokens=400`, 3 seeds per cell, median tok/s, `top_p 0.95 top_k 20`, ctx 32768, build `6d05498`. Speedup is against the no-drafter baseline **at the same temperature**:

翻訳待ち・原文を表示

テキスト生成

入力テキスト→ 出力テキスト

料金未確認マルチモーダル・チャネル料金は個別に確認

詳細を見る ↗

比較に追加出典・文書 ↗

Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Myric

基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。

基本情報を収録・詳細は順次追加

テキスト生成

入力テキスト→ 出力テキスト

料金未確認マルチモーダル・チャネル料金は個別に確認

詳細を見る ↗

比較に追加出典・文書 ↗

Kolibri-1-APEX-GGUF

Myric

基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。

テキスト生成

入力テキスト→ 出力テキスト

料金未確認マルチモーダル・チャネル料金は個別に確認

詳細を見る ↗

比較に追加出典・文書 ↗

Laguna-S-2.1-APEX-GGUF

Myric

MoE-aware, mixed-precision **APEX** quantization of poolside/Laguna-S-2.1 — a 118B-total / ~8B-active open-weight **agentic coding** MoE (48 layers, 256 routed + 1 shared expert, top-10 routing, GQA with per-head softplus output gating, mixed sliding-window/global attention, 1M context). APEX assigns quantization precision **per tensor role and per layer** instead of uniformly — keeping always-active shared experts and edge layers high-precision while compressing the sparse routed-expert middle layers hard. On a 256-expert MoE this is a large win: near-full-precision quality at a fraction of the size, and small enough to actually run this model locally. Paired here with the **DFlash** speculative-decoding draft model for fast generation.

翻訳待ち・原文を表示

テキスト生成

入力テキスト→ 出力テキスト

料金未確認マルチモーダル・チャネル料金は個別に確認

詳細を見る ↗

比較に追加出典・文書 ↗