One llama.cpp flag unlocks +33-39% decode speed for Qwen3.8-27B on consumer GPUs. The MTP head already ships inside your GGUF. Recipe, paired benchmarks, probe tool.
☆367Sep 26, 2026Updated last month
Alternatives and similar repositories for qwen38-mtp
Users that are interested in qwen38-mtp are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Deployment kit: Qwen3.8-27B EXL3 3.5bpw target + DFlash2 EXL3 5.0bpw speculative draft — launcher, config, model cards☆233Sep 1, 2026Updated 2 months ago
- Qwen3.8-27B on 16-32 GB Nvidia GPUs one-click install for Windows / Linux☆600Oct 9, 2026Updated 3 weeks ago
- ☆20Sep 17, 2026Updated last month
- A pixel art space shooter built entirely by a 9B AI model on a single RTX 3060. Zero hand-written code.☆106Mar 19, 2026Updated 7 months ago
- Local AI registry: one validated recipe per machine, with the evidence attached☆290Updated this week
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- run a 27b local ai on your gaming gpu, rtx 3060 and up, radeon rx 6600 and up in beta, 8gb or more, linux and windows, one line to instal…☆171Oct 17, 2026Updated 2 weeks ago
- OpenAI-compatible inference server for 4-bit Qwen3.8-27B on one RTX 3090, 262K context☆34Oct 29, 2026Updated last week
- Serve the official Ling-3.0-flash INT4 on one NVIDIA DGX Spark, 38.7 tok/s, faster than the community GGUF. Recipe, watchdog, benchmarks.☆26Aug 29, 2026Updated 2 months ago
- Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when …☆2,580Updated this week
- Qwen3.8-27B on SGLang for NVIDIA RTX PRO 6000 (96GB, SM120): DSpark speculative decoding, full 256K context, 8 concurrent requests☆61Aug 24, 2026Updated 2 months ago
- GLM-5.3-Flash EXL3 on one 24 GB RTX 3090 + DDR4: elastic GPU expert cache, zero-copy experts, AVX2 CPU tier, OpenAI API☆349Oct 18, 2026Updated 2 weeks ago
- Model-agnostic LLM benchmarking + tuning in one command. A model is a 15-key profile file. Built in production on a 4x NVIDIA DGX Spark c…☆10Sep 16, 2026Updated last month
- Fast, capable GLM-5.3-Flash FP4-Local on two DGX Sparks: opinionated Atlas serving and reproducible inference research.☆72Updated this week
- Three-DGX-Spark TP3 serving recipe for GLM-5.3 Flash☆21Oct 5, 2026Updated last month
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative…☆11Oct 13, 2026Updated 3 weeks ago
- Connect AI agents to 1,400+ APIs across search, social, SEO, finance, and more. Automatic tool discovery, unified billing, zero subscript…☆13Oct 27, 2026Updated last week
- NVIDIA Nemotron 3.5 Lightning 30B-A3B NVFP4 + DSpark for DGX Spark, RTX 5090/6000 PRO (SGLang)☆24Aug 11, 2026Updated 2 months ago
- Raw speed benchmark traces: Qwen3.8-27B on 4x RTX 3090. llama.cpp GGUF sweeps + 288-point vLLM vs SGLang, NVFP4 vs FP8 matrix.☆14Sep 28, 2026Updated last month
- Qwen3.8-27B EXL3 (4.00 bpw) + DFlash2 speculative decoding for ExLlamaV3, validated at 262k context on a 24 GB RTX 3090☆25Sep 20, 2026Updated last month
- Time-boxed passwordless sudo for agents. Grants expire on a systemd timer, validated with visudo before install, audited with sudoreplay.☆27Sep 3, 2026Updated 2 months ago
- Linux anonymity / OPSEC toolkit — Tor-by-default with a kill switch, leak verification, and pluggable upstreams. Authorized use only.☆24Sep 2, 2026Updated 2 months ago
- DGX Spark (GB10/sm_121) port: DeepSeek-V4-Flash 159B on one Spark, GLM-5.2 753B across two. Prepacked 2-bit planes + unified-memory fixes…☆19Jul 20, 2026Updated 3 months ago
- Validated vLLM recipe: Qwen3.8-27B-NVFP4 at full 262,144-token context on a single RTX 5090, GPU-only (no CPU offload).☆16Aug 20, 2026Updated 2 months ago
- Wordpress hosting with auto-scaling - Free Trial Offer • AdFully Managed hosting for WordPress and WooCommerce businesses that need reliable, auto-scalable performance. Cloudways SafeUpdates now available.
- GLM-5.2 QuantTrio TP=4+DCP2 on 4x NVIDIA DGX Spark (GB10) — tuned and measured under a real agent workload. Adaptive MTP, the indexer law…☆20Aug 24, 2026Updated 2 months ago
- Mixed-capability LLM benchmark for DGX Spark — 57 scenarios, 10 domains, partial-credit grading, trial statistics☆172Oct 13, 2026Updated 3 weeks ago
- qwen38-flash-next recipe for tp=2 GB10☆133Oct 29, 2026Updated last week
- ☆50Updated this week
- Truthful scheduler-step and streaming visualization for local LLM inference☆35Sep 6, 2026Updated 2 months ago
- Reproducible kit to deploy DeepSeek-V4-Flash-DSpark on a 2× NVIDIA DGX Spark (GB10) cluster: vLLM TP=2 over QSFP 200GbE, NVFP4 KV, DSpark…☆13Aug 24, 2026Updated 2 months ago
- Experimental llama.cpp fork for inference research and development☆900Updated this week
- An experimental and feature-rich inference engine built on top of llama.cpp☆83Updated this week
- ROCmFPX Family for AMD Hardware and Processors. More quants and special agent quants☆432Oct 22, 2026Updated 2 weeks ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- Fermion Research fork of llama.cpp: FV5/FV5B five-value ternary types for Neutrino GGUF packs (CPU + CUDA)☆14Sep 29, 2026Updated last month
- ☆780Updated this week
- What actually runs on an RTX 3060 12GB, if you have any card with 12GB vram, this is for you as well. measured, with commands☆163Aug 22, 2026Updated 2 months ago
- ReAlign models instantly, on demand. Per-request alignment for verified users and channels, or Always On. Use directly or with Hermes, Op…☆123Updated this week
- Fastest validated sglang recipe for Qwen3.8-27B (AWQ-INT4 + DSpark) on RTX 3090s — 1×/2×/4× GPU presets, image+video, full benchmarks☆64Oct 17, 2026Updated 3 weeks ago
- DeepSeek-V4-Flash-DSpark abliterated (uncensored) · ~100% refusal bypass · C1 ~57 tok/s · 1M ctx · 2× DGX Spark · HF weights☆57Aug 17, 2026Updated 2 months ago
- Mirai S (Qwen3.8-27B-S trellis GGUF) on a 12 GB RTX 4070: 262k q8_0 tiered cache, MTP drafting at every depth, harness-proofing, the laye…☆109Oct 26, 2026Updated last week