mlx-serve pack of Qwen/Qwen3.8-Flash-Next, the Qwen4 preview architecture (`model_type: qwen4_exp`), quantized to 3.3 bits per weight with imatrix calibration so it fits a 64 GB Mac with room for 64k of context. 52 GB resident. Includes the MTP head and the vision tower (image and video input). Measured against the bf16 checkpoint on 1200 held-out next-token positions (agent, code, prose, math; documents the calibration never saw). Top-1 is how often the pack's greedy token equals bf16's; mass is the bf16 probability of the pack's pick, with bf16's own pick as the ceiling. Same positions, same server, cold prefills. The 3.3 bpw pack keeps 99% of the 4-8 pack's probability mass at 21 GB less.
번역 대기 중 · 원문 표시
입력 텍스트→ 출력 텍스트
mlx-serve pack of Qwen/Qwen3.8-Flash-Next, the Qwen4 preview architecture (`model_type: qwen4_exp`). Runs on a 128 GB Mac with about 75 GB resident. Includes the MTP head and the vision tower (image and video input). This is the calibrated twin of Qwen3.8-Flash-Next-MLX-Serve-mixed-4-8bit: the same layout and the same size, byte for byte (every expert 4-bit group 64, the rest 8-bit), but every expert is quantized with an importance matrix collected on the bf16 model, so the scales follow the activations the weights actually see. Same speed, same memory, closer to bf16. The importance matrix was collected on the bf16 checkpoint over 1.35 M tokens of agent transcripts, code (including SWE-bench Lite issues and patches), prose and math, every expert of every layer hit. Quality is scored against the bf16 model's own logits on 290 held-out positions (top-1: the pack's greedy token equals bf16's; KLD: KL(bf16 || pack) over bf16's top-1024 tokens; the KLD mean is dominated by a few flat prose positions, so the median is the steadier number):
번역 대기 중 · 원문 표시
입력 텍스트→ 출력 텍스트
> **Use Qwen3.8-Flash-Next-MLX-Serve-iQ-MLX-4.7bpw instead.** > It is this pack re-quantized with an importance matrix collected on the bf16 > model: same layout, same 75 GB, same bytes read per token, closer to bf16 > (KLD median 0.0052 vs 0.0066, top-1 88.6% vs 87.9% against bf16's own > logits). This pack stays up for existing installs but gets no more updates. > > Switching from this pack? `ngram_table.bin` (32 GB) is identical in both. > Move it into the new pack's folder before downloading and the download > skips it: > > mlx-serve pack of Qwen/Qwen3.8-Flash-Next, the Qwen4 preview architecture (`model_type: qwen4_exp`). Runs on a 128 GB Mac with about 75 GB resident. Includes the MTP head and the vision tower (image and video input). Measured on an M4 Max 128 GB with mlx-serve 26.8.11: ~75 GB resident, decode ~60 tok/s serial and 78 tok/s with MTP (`--mtp`; +41% on code, a few percent slower on prose), prefill ~730 tok/s, a needle at 24.8k tokens recovered with sparse attention engaged. Prefix cache on, images included.
번역 대기 중 · 원문 표시
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
1. Read the ConjectureBench paper (arXiv:2510.11986) — nobody had implemented it locally 2. Built the first MLX-native LEAN-FIRE pipeline on Apple Silicon 3. Fine-tuned the first conjecture generation model that exists 4. Verified Dmitry Rybin's breaking counterexample to the 30-year-old Dinitz-Garg-Goemans conjecture within hours The core insight: **conjecture generation IS prediction.** Same pipeline whether you're breaking a 30-year-old math conjecture or sizing a Polymarket position. **The killer app.** Conjecture = Prediction. Formalization = Resolution criteria. Counterexample = The trade.
번역 대기 중 · 원문 표시
입력 텍스트→ 출력 텍스트
> **The most comprehensive open-source security agent model — in GGUF.** Runs in Ollama, LM Studio, llama.cpp, vLLM, and any GGUF runtime. 51/51 LoRA tensors merged. Identical to the MLX version. **Both versions are identical** — same 51/51 LoRA tensors, same 745K+ training data, same 12 training rounds. **People are NOT getting the most out of local LLMs.** A 35B MoE at Q4_K_M gives dramatically better output than a 7B model at the SAME speed — because only 3B params activate per token.
번역 대기 중 · 원문 표시
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
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。
기본 정보 등록 · 상세 정보 보완 예정
입력 텍스트→ 출력 텍스트
