AI Model2026
Enables local use of a GGUF-quantized DeepSeek-V4-Flash-0731 via Unsloth Dynamic quantizations; provides a Q8 (162GB) lossless option and smaller Q4 variants for lower-memory inference and agentic scenarios using Unsloth tooling.
AI Infra2026
Runs Kimi K3 inference on CPU by streaming model weights from storage instead of keeping them all in RAM. The project reports 8.24 GB peak memory, with substantially slower token generation at low-memory settings.
AI Model2026
Provides a GGUF-quantized, llama.cpp-compatible build of LiquidAI's LFM2.5-2.6B for local CPU inference and offline deployment. Supports multilingual generation and long-context workflows; optimized for low-memory, on-device use.
AI Model2026
A 35B additive ternary-quantized variant of Mach-1 that aims to preserve most capabilities of its BF16 teacher while reducing compute and memory; shows ~95% mean retention across 12 benchmarks and competitive per-task parity on several evaluations.
AI Model2026
A 20B ternary-weight Mixture-of-Experts reasoning LLM optimized for on-device and low-memory inference—delivers high throughput (200+ tok/s on M4) and an extremely long 131k-context for math/logic benchmarks, but is a preview with limited agentic fine-tuning.
AI Model2026
An uncensored NVFP4-quantized text encoder for MiniMax-H3 video generation that fits on a single 16 GB GPU. Mixed-precision bake (mostly NVFP4, embedding left as INT8), preserves ConvRot rotation semantics, and includes the unrotate step required to avoid corrupted conditioning.
AI Model2026
A 27B-parameter causal language model with a native vision encoder for image/video+text understanding, long-horizon agentic tasks, and tunable thinking-mode reasoning. Native 262,144-token context (extensible to 1,000,000) and production-focused inference recipes.
AI Video2026
Provides ComfyUI-compatible pruned/curve-form LoRA conversions of the MiniMax‑H3 Turbo 4-step audio‑video generation preview, including further-trained ckpt500 EMA and non‑EMA variants and an example ComfyUI workflow for low-step experiments.
Tao Feng, Fangxu Yu +10University of Illinois Urbana-Champaign, University of Maryland, College Park +3
Frames LLM routing as a sequential decision process and introduces LLMRouter plus the xRouteBench benchmark to develop, evaluate, and deploy learned routing policies across heterogeneous LLMs, optimizing response quality versus inference cost.
AI Audio2026
Generates complete songs (up to five minutes) from lyrics and a music description, producing 32 kHz stereo WAV with expressive vocals and long-range musical structure. Uses hierarchical LLMs fused with flow-matching/Flow-VAE synthesis for coherent arrangement and timbre; requires CUDA and integrates with Diffusers and SGLang-Omni.
AI Model2026
A MoE causal large language model for long-horizon agents, coding, and multi-step reasoning: 2.4T parameters (95B activated), native 262,144-token context (extensible to 1,010,000), multi-token prediction, and configurable thinking-mode reasoning controls.
AI Model2026
FP8-quantized checkpoint of the Qwen3.8 text-only causal LLM (2.4T params, 95B activated) for text-generation; preserves near-original performance, supports very long contexts (262k–1M), Mixture-of-Experts architecture, and is compatible with vLLM/SGLang/TokenSpeed. Thinking mode and preserve_thinking are enabled by default.