Strata fork: Qwen3.8-Flash-Next 125B MoE in NVFP4 on one RTX 20-50 card (12 GB+) and 64 GB of RAM or more. Setup installs our GPTQ quants (huihui abliterated, OrcaRouter) from Hugging Face. CPU experts on AVX-512 or AVX2 (+AVX-VNNI), all experts in RAM from 92 GB, a low-RAM mode below, vision, W4A8 prompts. Windows / CUDA 13.
☆224Oct 24, 2026Updated last week
Alternatives and similar repositories for strata-nvfp4
Users that are interested in strata-nvfp4 are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Local inference engine for Qwen3.8-Flash-Next on NVIDIA Blackwell☆694Updated this week
- C++/CUDA single-GPU inference engine for Qwen3.8 (incl. Qwen3.8-Flash-Next with offloaded experts) on the RTX 5090; grew from NInfer☆275Updated this week
- Qwen3.8-Flash-Next (MoE) and Qwen3.8-27B on one or two 16 GB RTX 50-series GPUs: experts offloaded to CPU, two-GPU tensor parallel. 64 GB…☆134Updated this week
- High-performance single-GPU inference for selected model checkpoints and GPUs.☆3,269Updated this week
- Xolotl (formerly Strata-DSV4.1): run DeepSeek-V4.1-Flash, all 552B, as a local coding agent on one gaming PC (RTX 4090 + 90 GB RAM) at up…☆467Updated this week
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Run GLM-5.3-Flash (321B), DeepSeek-V4.1-Flash (748B) and MiniMax-M3.1-Flash (428B) on your own GPUs - one engine, one dashboard, OpenAI- …☆2,585Updated this week
- Strata fork that runs GLM-5.3-Flash (320B-A18B, 157 GB) on one PC: RTX 5090 + 128 GB RAM + NVMe, ~30 tok/s chat. Fork of Niko1221/Strata …☆37Oct 13, 2026Updated 3 weeks ago
- High-performance single-GPU inference for selected model checkpoints and GPUs.☆186Sep 30, 2026Updated last month
- Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on loca…☆40,925Updated this week
- Async PLE SSD offload for Qwen3.8 Flash Next on a 64 GiB SM80 GPU: vLLM patch, setup guide, and measured PP/TG results.☆23Sep 24, 2026Updated last month
- High-performance single-GPU inference for selected model checkpoints and GPUs.☆24Oct 13, 2026Updated 3 weeks ago
- Linux P2P unlock and performance fixes for NVIDIA CMP 170HX☆39Sep 22, 2026Updated last month
- llama.cpp fork for significantly improved performance on Ampere generation Nvidia cards (especially RTX 3090 / 3090 Ti): example: 110+ to…☆223Updated this week
- Cross-tier CPU/GPU hybrid MoE inference for SGLang (plugin, tracks upstream main)☆64Updated 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.
- Unlock P2P comms between consumer NVIDIA GPUs☆697Oct 26, 2026Updated last week
- RTX 6000 96GB/Spark 128GB vLLM Recipe for Qwen 3.8 Flash Next☆23Sep 8, 2026Updated last month
- high-performance linear attention kernel library built on TileLang☆15Jun 8, 2026Updated 4 months ago
- 1Cat Studio: unified vLLM workbench, Codex Agent and creative canvas☆15Updated this week
- 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
- One line of NInfer for the Blackwell, Ada Lovelace, Ampere, Turing and Volta NVIDIA GPUs, consolidated from the forks that carry it: GGUF…☆196Updated this week
- LLM inference in C/C++☆34Aug 1, 2026Updated 3 months ago
- Run large AI models on your own PC: Qwen3.8-Flash-Next on Strata's engine, and GLM 5.3 Flash, DeepSeek V4 Flash, LFM2.5, Qwen3.6, Ornith …☆887Updated this week
- Private Strata customizations for 9950x3d2+dual 7900xtx; upstream Niko1221/Strata☆70Updated this week
- Proton VPN Special Offer - Get 70% off • AdSpecial partner offer. Trusted by over 100 million users worldwide. Tested, Approved and Recommended by Experts.
- Qwen3.8-Flash-Next 180B NVFP4 @ 231 tok/s on one RTX PRO 6000 Blackwell (sm120) - sglang patches, launchers, full optimization analysis☆30Aug 31, 2026Updated 2 months ago
- Persistent project-scoped knowledge base for Claude Code and Codex: quick and deep research depths, contrarian pass, source-independence …☆15Jul 23, 2026Updated 3 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-Flash-Next (125B MoE) on a 8GB+ NVIDIA GPU: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API …☆89Oct 9, 2026Updated 3 weeks ago
- ENGRAFT: write facts into an LLM's n-gram (Engram) memory table as a removable overlay: 100 facts in a 9 MB file, no weights touched, ver…☆60Sep 30, 2026Updated last month
- ☆31Sep 30, 2026Updated last month
- KVarN, KV cache precision tail, low-bit quants in llama.cpp for longer context of better precision in the same VRAM☆1,183Updated this week
- Local LLM Inference Speed Test Tool☆249Oct 21, 2026Updated 2 weeks ago
- Qwen3.8-Flash-Next on 2× RTX 3090: with 128 GB RAM up to 4,191 tok/s prefill · 111.5 tok/s decode (131K prompt), full 256K window at 2,86…☆134Oct 15, 2026Updated 3 weeks ago
- 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 GTK GUI to PROM/GAL/Microcontroller Burner hardware. Supported Burners: JTAG Altera byteblaster, Funprog, GALBlast, noname JTAG, STK200…☆13Jun 9, 2017Updated 9 years ago
- vLLM 0.33 fork for consumer GPUs: 3 x 262K sessions on 4x RTX 3090, Qwen 3.8 Flash-Next at 207 tok/s, up to 17 on bigger cards. Seven che…☆45Updated this week
- Minimax H3 prompt☆12Aug 11, 2026Updated 2 months ago
- A tool to unlobotomize your NVIDIA card!☆834Oct 28, 2026Updated last week
- GLM-5.3-Flash EXL3 2.05 on one DGX Spark — TP1 vLLM, DFlash2 K5, concurrency, 262K context, and published fidelity measurements☆57Sep 22, 2026Updated last month
- Tensor-parallel DeepSeek V4 Flash inference on dual AMD Strix Halo over OdinLink USB4/TB5 RDMA or Mellanox RoCE v2.☆23Sep 20, 2026Updated last month
- Frontier MoE models on RTX PRO 6000 / RTX 5090 (sm_120): DeepSeek-V4.1-Flash on 4x RTX PRO 6000, as a GPU+CPU hybrid on 4x RTX 5090, and …☆568Updated this week