exl3

Kyojin: the Yamz inference engine for AMD Strix Halo (ROCm, gfx1151), built on ExLlamaV3. Runs 300B-class MoE models on one 128 GB mini PC.

DeepSeek-V4.1-Flash on two DGX Sparks: EXL3 routed experts, 262k context, 2M-token KV, vision, tools, MTP, bundled with Pi

GLM-5.3-Flash EXL3 (320B MoE) on 2x NVIDIA DGX Spark — production serving kit, 1M context, 97%+ multi-session prefix caching, DFlash2 spec decode

Validated GLM-5.3 Flash recipe for 2x NVIDIA RTX PRO 6000 Blackwell 96GB: 262K context, EXL3/TR3, adaptive MTP, tools, and vision.

Fast, capable GLM-5.3-Flash NVFP4 on two DGX Sparks: opinionated Atlas serving and reproducible inference research.

A tiered-memory system design for workloads that don't fit in RAM: measure the working set, pin the hot tier, stream the cold tier from flash. Ships the residency calculator, measurement harnesses, and the build recipes behind it. Predictions validated against public benchmarks.

Serve EXL3 (ExLlamaV3 trellis) quantized models on vLLM fork runtimes — any architecture, mixed per-layer bitrates, composable with source-format non-routed weights

vLLM recipe: GLM-5.3-Flash EXL3 K2 on one DGX Spark GB10. Native MTP k=2. Measured tok/s.

Qwen3.8-27B EXL3 (4.00 bpw) + DFlash2 speculative decoding for ExLlamaV3, validated at 262k context on a 24 GB RTX 3090

GLM-5.3-Flash EXL3 on 4x RTX PRO 6000 with TensorFold: one-command recipe and engine patches. By Aevonix Research in collaboration with Mia's AI Lab.

Qwen3.8-Flash-Next EXL3 3.05bpw on ONE DGX Spark (GB10): ~2.4x prefill (700->1777 t/s), prompt-lookup drafter, native vision/video, 262k ctx. Unofficial TensorFold v0.6.5 fork, 18 patches, byte-exact receipts.

Reproducible 262K GLM-5.3 Flash Uncensored EXL3+DFlash2 deployment for 2x RTX PRO 6000 Blackwell

Native EXL3 inference engine for macOS and Apple Silicon. Rust, MLX and Metal runtime, with a SwiftUI app and CLI.

GLM-5.3-Flash EXL3 on 2x RTX PRO 6000 with TensorFold: one-command recipe and engine patches. By Aevonix Research in collaboration with Mia's AI Lab.

High-performance runtime extensions for vLLM.

GLM-5.3 EXL3 on 4x RTX PRO 6000 with TensorFold: one-command recipe and engine patches. By Aevonix Research in collaboration with Mia's AI Lab.

DeepSeek-V4-Flash-Vision-Exp (EXL3 MixedK, 256 experts, uncensored) on one NVIDIA DGX Spark with vLLM + sparkinfer: 245,760 context, vision + DSpark speculative decoding, CUDA graphs. Recipe, overlay patches, benchmarks, receipts.

GLM-5.3-Flash EXL3 on three DGX Sparks (TP3/EP3, 1M context, vision, DFlash2) — jspark3 fork, traps, measurements

Production-tuned deployment recipe: GLM-5.3-Flash 320B (EXL3 4-bit) on 2x NVIDIA DGX Spark - 850K context, DFlash2 + adaptive-k speculative decoding, full .env tuning and pitfalls

LLM inference server for ExLlamaV3 / EXL3, with OpenAI- and Anthropic-compatible APIs optimized for Agent workloads.