Filter by type:
Filter by tags:
qwen3.8-2b-q4
Qwen3.8-2B is Empero AI's smallest Qwen3.8 reasoning distillation. It uses
the Qwen3.5-2B architecture and targets mathematics, instruction
following, tool use, and edge deployment with a native 262K-token context
window. This default entry uses Q4_K_M weights; a higher-quality Q8_0
build is available as a variant.
Repository: localaiLicense: apache-2.0
nemotron-3.5-lightning-30b-a3b-q4
NVIDIA Nemotron 3.5 Lightning is a text-only hybrid Mamba-2, attention, and
mixture-of-experts model with 30B total parameters and 3B active parameters.
It targets reasoning, coding, tool use, multilingual chat, and long-context
agent workflows, with a context window of up to one million tokens.
This entry uses the official Q4_K_M GGUF. Automatic variant selection can
choose the smaller NVFP4 build or the higher-quality Q8_0 build when it fits.
Repository: localaiLicense: openmdw-1.1
nemotron-3.5-lightning-30b-a3b-nvfp4
NVIDIA Nemotron 3.5 Lightning 30B-A3B in the official NVFP4 GGUF format.
This is the smallest linked build and retains the model's reasoning, coding,
tool-use, multilingual, and long-context capabilities.
Repository: localaiLicense: openmdw-1.1
muse-glimmer-30b
Muse Glimmer is Meta Superintelligence Labs' Apache-2.0 dense 30B model for
autonomous agentic work, coding, tool use, long-horizon reasoning, and
multimodal understanding. It supports more than 100 languages, interleaved
text and image input through its 1.8B-parameter perception encoder, and a
131K-token context window.
This entry uses the publisher's higher-quality dynamic K-quant GGUF and
official quantized vision projector. Automatic variant selection can use
the smaller 17 GB quantization or a DFlash-accelerated build when it fits.
Repository: localaiLicense: apache-2.0
muse-glimmer-30b-17gb
Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
perception encoder. It preserves the model's agentic, coding, tool-use,
multilingual, and image-understanding capabilities for hosts with less
memory than the dynamic quantization requires.
Repository: localaiLicense: apache-2.0
muse-glimmer-30b-17gb-dflash
Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
perception encoder and DFlash drafter. This is the lowest-memory published
build that retains image understanding and block-speculative decoding.
Flash attention is enabled for the DFlash path.
Repository: localaiLicense: apache-2.0
deepseek-v4-flash-0731
# DeepSeek-V4-Flash-0731
Technical Report👁️
## Introduction
**DeepSeek-V4-Flash-0731** is the official release of **DeepSeek-V4-Flash**, superseding the preview version, with substantially enhanced agentic capabilities. It has the same model structure as DeepSeek-V4-Flash-DSpark, i.e. it comes with a speculative decoding module attached.
DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on benchmarks listed below despite its far smaller activated parameter count, and is broadly competitive with the strongest proprietary models available.
Notes:
1. For the Code Agent tasks among the public benchmarks above, DeepSeek-V4-Flash-0731 is evaluated with the minimal mode of DeepSeek Harness (to be released) as the agent framework, using the `max` reasoning effort level with `temperature = 1.0, top_p = 0.95`.
2. † DSBench-FullStack is an internal full-stack development test set; DSBench-Hard is an internal test set of difficult coding-agent problems.
## Chat Template
...
Repository: localaiLicense: mit
parable-granite-4.1-3b-claude-fable-5
# Parable-Granite-4.1-3B-Claude-Fable-5
Granite 4.1 3B fine-tuned on genuine Claude Fable 5 and GPT-5.5 agent
traces (planning, tool use, reasoning from real agent sessions).
Agent-flavored small model: terminal workflows, idiomatic code fixes,
explanations. v2 recipe: completion-masked SFT, replay mix, seed-averaged
weights. Published corpus and eval harness.
Repository: localaiLicense: apache-2.0
pocket-35b-q2
POCKET-35B is an Apache-2.0 Qwen3.5-family mixture-of-experts model from
FINAL-Bench/VIDRAFT, derived from Darwin-36B-Opus and packaged for stock
llama.cpp. This entry uses the smaller Q2_K GGUF quantization.
Repository: localaiLicense: apache-2.0
pocket-26b-q2
POCKET-26B is an Apache-2.0 Gemma 4 26B-A4B mixture-of-experts model from
FINAL-Bench/VIDRAFT, tuned for Korean and packaged for stock llama.cpp.
This entry uses the smaller Q2_K GGUF quantization.
Repository: localaiLicense: apache-2.0
qwen3.6-35b-a3b-genesis-hermes-v7-apex-compact
Qwen3.6-35B-A3B Genesis Hermes V7 in the smaller APEX Compact GGUF format,
with the shared F16 multimodal projector. This build preserves the model's
multimodal, reasoning, coding, and agentic capabilities for hosts with less
memory than the recommended full APEX build.
Repository: localaiLicense: apache-2.0
qwen3.6-35b-a3b-genesis-hermes-v7-mtp-apex-compact
Qwen3.6-35B-A3B Genesis Hermes V7 in the smaller APEX Compact GGUF format
with native multi-token prediction enabled for speculative decoding, plus
the shared F16 multimodal projector.
Repository: localaiLicense: apache-2.0
inkling-small
Inkling Small is a 276B-parameter mixture-of-experts multimodal model with 12B active parameters for text, image, and audio understanding, instruction following, coding, and tool use. This entry uses the Q4_K_M GGUF quantization, whose five language-model shards total approximately 162.5 GB.
Repository: localaiLicense: apache-2.0
inkling-small-iq2-m
Inkling Small is a 276B-parameter mixture-of-experts multimodal model with 12B active parameters for text, image, and audio understanding, instruction following, coding, and tool use. This entry uses the IQ2_M GGUF quantization, whose three language-model shards total approximately 82.4 GB.
Repository: localaiLicense: apache-2.0
bonsai-8b-1bit
Bonsai 8B (PrismML) is an end-to-end 1-bit language model built on the Qwen3-8B dense
architecture (GQA, SwiGLU, RoPE, RMSNorm, 36 layers, 65,536 context). Every weight is a
single sign bit (`-scale` / `+scale`) with one FP16 scale per group of 128 weights, for
an effective 1.125 bits/weight and a ~1.15 GB footprint (14.2x smaller than FP16) while
matching full-precision 8B instruct models at ~70.5 average across 6 benchmark
categories.
The Q1_0 quantization is only decodable by the PrismML llama.cpp fork, so this entry
runs on LocalAI's `bonsai` backend (that fork), not the stock `llama-cpp` backend.
License: Apache 2.0.
Repository: localaiLicense: apache-2.0
ternary-bonsai-8b
Ternary Bonsai 8B (PrismML) is a 1.58-bit ternary language model on the Qwen3-8B dense
architecture. Each weight takes a value from {-1, 0, +1} with one shared FP16 scale per
group of 128 weights (GGUF Q2_0, ~2.18 GB deployed, 7.5x smaller than FP16). The extra
zero state recovers more of the full-precision model than the 1-bit build: it ranks 2nd
among compared 6-9B models at 75.5 average despite being ~1/8th their size.
Q2_0 is the recommended, ternary-lossless variant. The Q2_0 kernels are only in the
PrismML llama.cpp fork, so this runs on LocalAI's `bonsai` backend. License: Apache 2.0.
Repository: localaiLicense: apache-2.0
bonsai-27b-1bit
Bonsai 27B (PrismML) is a full 27B-class reasoning model in end-to-end 1-bit weights,
derived from the Qwen3.6-27B hybrid-attention backbone (~75% linear attention, 262K
context). At a true 1.125 bits/weight it deploys in ~3.9 GB (~14.2x smaller than FP16)
while retaining 89.5% of FP16 intelligence across 15 thinking-mode benchmarks (math
91.66, coding 81.88). Ships an optional 4-bit vision tower (mmproj) for image input,
included here.
The Q1_0_g128 weights and hybrid-attention kernels are only in the PrismML llama.cpp
fork, so this runs on LocalAI's `bonsai` backend. A GPU is recommended. License: Apache 2.0.
Repository: localaiLicense: apache-2.0
laguna-xs-2.1-apex-i-mini
Laguna XS 2.1 in the 12.8 GB APEX-I Mini format, the smallest importance-matrix APEX build for llama.cpp.
License: OpenMDW 1.1.
Repository: localaiLicense: other
qwen3.6-35b-a3b-dflash
Qwen3.6-35B-A3B (Mixture-of-Experts, ~3B active per token) paired with its DFlash block-diffusion drafter for speculative decoding on the llama.cpp backend. DFlash speedups on MoE targets are smaller than on dense models, but still useful.
DFlash produces a whole block of draft tokens in a single forward pass and injects the target model's hidden states into the drafter's attention, which keeps the drafter tiny while making drafting GPU-friendly. The UD-Q4_K_M file carries the full Qwen3.6-35B-A3B target; the ~0.4 GB Q8_0 drafter (`draft-dflash`) accelerates generation without changing the target's outputs. The drafter is not a standalone chat model: it only runs paired with the target, which is why both are bundled here.
Flash attention is required for DFlash and is enabled in this config. A GPU is recommended.
License: Apache 2.0.
Repository: localaiLicense: apache-2.0
grug-12b
Grug 12B is kai-os's compact-reasoning fine-tune of Gemma 4 12B IT. It
targets shorter, denser reasoning traces while preserving constraints,
branching decisions, edge cases, and final-answer checks. This entry uses
Bartowski's Q4_K_M quantization and includes the multimodal projector for
Gemma 4 image inputs.
The model is experimental and its reported evaluation is a small local
math proxy rather than a broad benchmark. Review the upstream model card's
dataset provenance and `other` license before commercial or sensitive use.
Repository: localaiLicense: other
grug-12b-q8
Grug 12B is kai-os's compact-reasoning fine-tune of Gemma 4 12B IT. This
high-quality variant uses Bartowski's Q8_0 quantization and includes the
multimodal projector for Gemma 4 image inputs.
The model is experimental and its reported evaluation is a small local
math proxy rather than a broad benchmark. Review the upstream model card's
dataset provenance and `other` license before commercial or sensitive use.
Repository: localaiLicense: other
Page 1