Filter by type:
Filter by tags:
qwen3.5-9b-defiant-fable
Qwen3.5 9B Defiant Fable in the plain NEO-imatrix Q4_K_M GGUF format.
This fallback offers the same multimodal reasoning, coding, and creative
capabilities without enabling multi-token prediction.
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
glm-5.2
# GLM-5.2
👋 Join our WeChat or Discord community.
📖 Check out the GLM-5.2 blog and GLM-5 Technical report.
📍 Use GLM-5.2 API services on Z.ai API Platform.
🔜 Try GLM-5.2 here.
[Paper]
[GitHub]
## Introduction
We're introducing GLM-5.2, our latest flagship model for long-horizon tasks. It marks a substantial leap in long-horizon task capability over its predecessor GLM-5.1 and, for the first time, delivers that capability on a **solid 1M-token context**. GLM-5.2's new capabilities include:
- **Solid 1M Context:** A solid 1M-token context that stably sustains long-horizon work
- **Advanced Coding with Flexible Effort**: Stronger coding capabilities with multiple thinking effort levels to balance performance and latency
- **Improved Architecture**: We propose IndexShare, which reuses the same indexer across every four sparse attention layers, reducing per-token FLOPs by 2.9× at a 1M context length. We also improve GLM-5.2’s MTP layer for speculative decoding, increasing the acceptance length by up to 20%
- **Pure Open**: An MIT open-source license — no regional limits, technical access without borders
## Benchmark
## Serve GLM-5.2 Locally
...
Repository: localaiLicense: mit
gemma-4-12b-agentic-fable5-composer2.5-v2-3.5x-tau2
Hugging Face |
GitHub |
Launch Blog |
Documentation
License: Apache 2.0 | Authors: Google DeepMind
> [!Note]
> This model card is for the Gemma 4 12B Unified model, which is part of the Gemma 4 family of open models. Built with the same multimodal functionality as Gemma 4 E2B and E4B (text, audio, image, and video inputs), it brings native audio and vision understanding directly to local environments without the need for separate encoders. This unified approach to multimodality makes the model encoder-free, offering a deployment size that is perfect for consumer devices and streamlined local execution.
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants. Gemma 4 features a context window of up to 256K tokens and maintains multilingual support in over 140 languages.
...
Repository: localaiLicense: apache-2.0
gemma-4-12b-coder-fable5-composer2.5-v1
Hugging Face |
GitHub |
Launch Blog |
Documentation
License: Apache 2.0 | Authors: Google DeepMind
> [!Note]
> This model card is for the Gemma 4 12B Unified model, which is part of the Gemma 4 family of open models. Built with the same multimodal functionality as Gemma 4 E2B and E4B (text, audio, image, and video inputs), it brings native audio and vision understanding directly to local environments without the need for separate encoders. This unified approach to multimodality makes the model encoder-free, offering a deployment size that is perfect for consumer devices and streamlined local execution.
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants. Gemma 4 features a context window of up to 256K tokens and maintains multilingual support in over 140 languages.
...
Repository: localaiLicense: gemma
qwen3.6-35b-a3b-claude-4.6-opus-reasoning-distilled
# 🔥 Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled
A reasoning SFT fine-tune of `Qwen/Qwen3.6-35B-A3B` on chain-of-thought (CoT) distillation mostly sourced from Claude Opus 4.6. The goal is to preserve Qwen3.6's strong agentic coding and reasoning base while nudging the model toward structured Claude Opus-style reasoning traces and more stable long-form problem solving.
The training path is text-only. The Qwen3.6 base architecture includes a vision encoder, but this fine-tuning run did not train on image or video examples.
- **Developed by:** @hesamation
- **Base model:** `Qwen/Qwen3.6-35B-A3B`
- **License:** apache-2.0
This fine-tuning run is inspired by Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled, including the notebook/training workflow style and Claude Opus reasoning-distillation direction.
[](https://x.com/Hesamation) [](https://discord.gg/vtJykN3t)
## Benchmark Results
The MMLU-Pro pass used 70 total questions per model: `--limit 5` across 14 MMLU-Pro subjects. Treat this as a smoke/comparative check, not a release-quality full benchmark.
...
Repository: localaiLicense: apache-2.0
nanbeige4.1-3b-q8
Nanbeige4.1-3B is built upon Nanbeige4-3B-Base and represents an enhanced iteration of our previous reasoning model, Nanbeige4-3B-Thinking-2511, achieved through further post-training optimization with supervised fine-tuning (SFT) and reinforcement learning (RL). As a highly competitive open-source model at a small parameter scale, Nanbeige4.1-3B illustrates that compact models can simultaneously achieve robust reasoning, preference alignment, and effective agentic behaviors.
Key features:
Strong Reasoning: Capable of solving complex, multi-step problems through sustained and coherent reasoning within a single forward pass, reliably producing correct answers on benchmarks like LiveCodeBench-Pro, IMO-Answer-Bench, and AIME 2026 I.
Robust Preference Alignment: Outperforms same-scale models (e.g., Qwen3-4B-2507, Nanbeige4-3B-2511) and larger models (e.g., Qwen3-30B-A3B, Qwen3-32B) on Arena-Hard-v2 and Multi-Challenge.
Agentic Capability: First general small model to natively support deep-search tasks and sustain complex problem-solving with >500 rounds of tool invocations; excels in benchmarks like xBench-DeepSearch (75), Browse-Comp (39), and others.
Repository: localaiLicense: apache-2.0
nanbeige4.1-3b-q4
Nanbeige4.1-3B is built upon Nanbeige4-3B-Base and represents an enhanced iteration of our previous reasoning model, Nanbeige4-3B-Thinking-2511, achieved through further post-training optimization with supervised fine-tuning (SFT) and reinforcement learning (RL). As a highly competitive open-source model at a small parameter scale, Nanbeige4.1-3B illustrates that compact models can simultaneously achieve robust reasoning, preference alignment, and effective agentic behaviors.
Key features:
Strong Reasoning: Capable of solving complex, multi-step problems through sustained and coherent reasoning within a single forward pass, reliably producing correct answers on benchmarks like LiveCodeBench-Pro, IMO-Answer-Bench, and AIME 2026 I.
Robust Preference Alignment: Outperforms same-scale models (e.g., Qwen3-4B-2507, Nanbeige4-3B-2511) and larger models (e.g., Qwen3-30B-A3B, Qwen3-32B) on Arena-Hard-v2 and Multi-Challenge.
Agentic Capability: First general small model to natively support deep-search tasks and sustain complex problem-solving with >500 rounds of tool invocations; excels in benchmarks like xBench-DeepSearch (75), Browse-Comp (39), and others.
Repository: localaiLicense: apache-2.0
silero-vad-sherpa
Silero VAD served through the sherpa-onnx backend. Uses the same ONNX weights as the dedicated silero-vad backend, loaded through sherpa-onnx's C VAD API. Pairs with the sherpa-onnx ASR entries for round-trip audio pipelines.
Repository: localaiLicense: mit
qwen3.6-27b-nvfp4-vllm-cpp
Qwen3.6-27B in NVFP4, served by vllm.cpp: LocalAI's own C++ port of vLLM, with
no Python at inference time. This is the reference text-generation checkpoint
the engine is gated on, token-for-token identical to vLLM's own greedy output
over the 235-prompt correctness battery, and measured at or above vLLM's
throughput at every concurrency from 1 to 32.
The weights are PINNED to revision 890bdef7. That pin is load-bearing, not
housekeeping: the same repository name was later re-quantized to FP8 W8A8
throughout, so an unpinned copy of this entry serves entirely different
weights with no error and none of the measured behaviour above.
Needs a Blackwell-class NVIDIA GPU (NVFP4 has no kernel on older
architectures) and roughly 25 GB of weights plus KV cache. Tool calling and
the thinking split are parsed inside the engine.
Repository: localaiLicense: apache-2.0
qwen3.6-27b-nvfp4-mtp-vllm-cpp
Qwen3.6-27B NVFP4 on vllm.cpp with MTP speculative decoding enabled.
MTP (Multi-Token Prediction) drafts from a head that ships inside the target
checkpoint's own mtp.* tensors, so there is no second model to download and
no extra weights to manage. The verifier accepts roughly 85% of drafted
tokens on prose and 92% on code, worth about 1.5x to 1.6x the decode
throughput of the same weights with speculation off, and it holds that lead
at concurrency 2, 4 and 8.
Same weights and same revision pin as qwen3.6-27b-nvfp4-vllm-cpp; install
that entry instead if you would rather not spend the extra memory. The
speculative state (a doubled recurrent-state slot plus the draft cache and
head) costs roughly 3.6 GB on top of the base footprint.
MTP here is depth 1 by construction: the engine refuses
num_speculative_tokens above 1 for this method.
Repository: localaiLicense: apache-2.0
qwen3.6-27b-nvfp4-dflash-vllm-cpp
Qwen3.6-27B NVFP4 on vllm.cpp with DFlash block-diffusion speculative
decoding: the fastest configuration of this model the engine ships.
Where MTP drafts one token at a time, DFlash drafts a whole 16-token block in
a single non-autoregressive pass from a separate 3.5 GB drafter, then the
target verifies the block in one step. At concurrency 1 that measures 2.9x
the throughput of the same weights with speculation off, and at or above
vLLM's own DFlash-on decode.
Both checkpoints are installed for you: the target as a revision-pinned
snapshot, the drafter into models/Qwen3.6-27B-DFlash, which is where the
backend looks when speculative_config.model names it. The drafter shares the
target's embed_tokens and lm_head, so the two are not independently
swappable.
Needs a Blackwell-class NVIDIA GPU and roughly 28 GB of weights in total.
Repository: localaiLicense: apache-2.0
minimax-h3-ref2va-q4
MiniMax-H3 served by vllm.cpp, LocalAI's own C++ port of vLLM. It generates
video AND audio jointly, so a clip comes back as an MP4 with a real
soundtrack rather than a silent render.
This is the Q4_K_M quantisation of the Ref2VA partition, the one that takes
REFERENCE conditioning: a reference image, a reference clip, or reference
audio, prepended as their own blocks so the subject or style carries into the
generated video. For plain text-to-video or first/last-frame conditioning use
minimax-h3-fl2va-q4 instead - the two partitions are separate checkpoints and
each refuses the other's tasks.
Use this Q4_K_M build, NOT the NVFP4 Ref2VA weights: NVFP4 renders a
multicolour patch grid, and it took three investigations upstream to
establish that the fault is the quantisation rather than the reference path.
On Q4_K_M the same code renders coherently.
Roughly 40 GB of weights across five files, plus the two VAE configs that
carry the latent statistics. The default canvas is 1344x768 at 124 frames and
24 fps, about 5.2 seconds. Generation is slow - roughly 176 s per denoise
step at that canvas on a 20-SM device, so the 50-step default is a multi-hour
job. Muxing the finished frames needs ffmpeg on the host.
Repository: localaiLicense: other
ace-step-turbo
ACE-Step 1.5 Turbo is a music generation model that can create music from text descriptions,
lyrics, or audio samples. Supports both simple text-to-music and advanced music generation
with metadata like BPM, key scale, and time signature.
Repository: localaiLicense: mit
speechbrain-ecapa-tdnn
Speaker (voice) recognition with SpeechBrain's ECAPA-TDNN trained
on VoxCeleb. 192-d L2-normalised embeddings, ~1.9% Equal Error
Rate on VoxCeleb1-O. APACHE 2.0 — commercial-safe.
The checkpoint is auto-downloaded from HuggingFace on first
LoadModel (no separate weight file in gallery `files:`). Points at
the upstream SpeechBrain HF repo directly — same bytes every
deployment.
Repository: localaiLicense: apache-2.0
rfdetr-cpp-nano
RF-DETR Nano object detection model, served via the native rfdetr.cpp backend (ggml + purego, no Python).
Q8_0 quantization is the recommended default for CPU: same accuracy as F16/F32, ~20MB on disk, fastest CPU latency.
Pure C++/ggml runtime; no Python dependencies. Drop-in for the /v1/detection endpoint.
Repository: localaiLicense: apache-2.0
edgetam
EdgeTAM is an ultra-efficient variant of the Segment Anything Model (SAM) for image segmentation.
It uses a RepViT backbone and is only ~16MB quantized (Q4_0), making it ideal for edge deployment.
Supports point-prompted and box-prompted image segmentation via the /v1/detection endpoint.
Powered by sam3.cpp (C/C++ with GGML).
Repository: localaiLicense: apache-2.0
fast-math-qwen3-14b
By applying SFT and GRPO on difficult math problems, we enhanced the performance of DeepSeek-R1-Distill-Qwen-14B and developed Fast-Math-R1-14B, which achieves approx. 30% faster inference on average, while maintaining accuracy.
In addition, we trained and open-sourced Fast-Math-Qwen3-14B, an efficiency-optimized version of Qwen3-14B`, following the same approach.
Compared to Qwen3-14B, this model enables approx. 65% faster inference on average, with minimal loss in performance.
Technical details can be found in our github repository.
Note: This model likely inherits the ability to perform inference in TIR mode from the original model. However, all of our experiments were conducted in CoT mode, and its performance in TIR mode has not been evaluated.
Repository: localaiLicense: apache-2.0
vulpecula-4b
**Vulpecula-4B** is fine-tuned based on the traces of **SK1.1**, consisting of the same 1,000 entries of the **DeepSeek thinking trajectory**, along with fine-tuning on **Fine-Tome 100k** and **Open Math Reasoning** datasets. This specialized 4B parameter model is designed for enhanced mathematical reasoning, logical problem-solving, and structured content generation, optimized for precision and step-by-step explanation.
Repository: localaiLicense: apache-2.0
qwen3-55b-a3b-total-recall-deep-40x
WARNING: MADNESS - UN HINGED and... NSFW. Vivid prose. INTENSE. Visceral Details. Violence. HORROR. GORE. Swearing. UNCENSORED... humor, romance, fun.
Qwen3-55B-A3B-TOTAL-RECALL-Deep-40X-GGUF
A highly experimental model ("tamer" versions below) based on Qwen3-30B-A3B (MOE, 128 experts, 8 activated), with Brainstorm 40X (by DavidAU - details at bottom of this page).
These modifications blow the model (V1) out to 87 layers, 1046 tensors and 55B parameters.
Note that some versions are smaller than this, with fewer layers/tensors and smaller parameter counts.
The adapter extensively alters performance, reasoning and output generation.
Exceptional changes in creative, prose and general performance.
Regens of the same prompt - even with the same settings - will be very different.
THREE example generations below - creative (generated with Q3_K_M, V1 model).
ONE example generation (#4) - non creative (generated with Q3_K_M, V1 model).
You can run this model on CPU and/or GPU due to unique model construction, size of experts and total activated experts at 3B parameters (8 experts), which translates into roughly almost 6B parameters in this version.
Two quants uploaded for testing: Q3_K_M, Q4_K_M
V3, V4 and V5 are also available in these two quants.
V2 and V6 in Q3_k_m only; as are: V 1.3, 1.4, 1.5, 1.7 and V7 (newest)
NOTE: V2 and up are from source model 2, V1 and 1.3,1.4,1.5,1.7 are from source model 1.
Repository: localaiLicense: apache-2.0
qwen3-42b-a3b-stranger-thoughts-deep20x-abliterated-uncensored-i1
WARNING: NSFW. Vivid prose. INTENSE. Visceral Details. Violence. HORROR. GORE. Swearing. UNCENSORED... humor, romance, fun.
Qwen3-42B-A3B-Stranger-Thoughts-Deep20x-Abliterated-Uncensored
This repo contains the full precision source code, in "safe tensors" format to generate GGUFs, GPTQ, EXL2, AWQ, HQQ and other formats. The source code can also be used directly.
ABOUT:
Qwen's excellent "Qwen3-30B-A3B", abliterated by "huihui-ai" then combined Brainstorm 20x (tech notes at bottom of the page) in a MOE (128 experts) at 42B parameters (up from 30B).
This pushes Qwen's abliterated/uncensored model to the absolute limit for creative use cases.
Prose (all), reasoning, thinking ... all will be very different from reg "Qwen 3s".
This model will generate horror, fiction, erotica, - you name it - in vivid, stark detail.
It will NOT hold back.
Likewise, regen(s) of the same prompt - even at the same settings - will create very different version(s) too.
See FOUR examples below.
Model retains full reasoning, and output generation of a Qwen3 MOE ; but has not been tested for "non-creative" use cases.
Model is set with Qwen's default config:
40 k context
8 of 128 experts activated.
Chatml OR Jinja Template (embedded)
IMPORTANT:
See usage guide / repo below to get the most out of this model, as settings are very specific.
USAGE GUIDE:
Please refer to this model card for
Specific usage, suggested settings, changing ACTIVE EXPERTS, templates, settings and the like:
How to maximize this model in "uncensored" form, with specific notes on "abliterated" models.
Rep pen / temp settings specific to getting the model to perform strongly.
https://huggingface.co/DavidAU/Qwen3-18B-A3B-Stranger-Thoughts-Abliterated-Uncensored-GGUF
GGUF / QUANTS / SPECIAL SHOUTOUT:
Special thanks to team Mradermacher for making the quants!
https://huggingface.co/mradermacher/Qwen3-42B-A3B-Stranger-Thoughts-Deep20x-Abliterated-Uncensored-GGUF
KNOWN ISSUES:
Model may "mis-capitalize" word(s) - lowercase, where uppercase should be - from time to time.
Model may add extra space from time to time before a word.
Incorrect template and/or settings will result in a drop in performance / poor performance.
Repository: localaiLicense: apache-2.0
Page 1