Filter by type:
Filter by tags:
inkling
# Inkling
BF16 |
NVFP4 |
Playground |
Tinker Cookbook |
Acceptable Use
## 1. General Information
Inkling is a general-purpose multimodal model that accepts text, image and audio inputs and generates text outputs. It is intended for use in English and other languages, and across multiple coding languages. The model is designed to be used by developers building AI-powered applications, including agentic and tool-use systems, coding assistants, chatbots, and retrieval-augmented generation systems, and is suitable for general-purpose conversational use, instruction-following, and other natural language and multimodal tasks. It is released with open weights to support research, fine-tuning and integration into third-party products by downstream developers.
**Languages:** English, with general multilingual capabilities across other languages.
## 2. Getting Started
Try Inkling on the Tinker Playground or access via API using the Tinker Cookbook.
Inkling supports local deployment using the following open-source libraries:
* SGLang (recipe, PR)
* vLLM (recipe, PR)
* TokenSpeed (recipe, PR)
* Unsloth (recipe, PR)
* Huggingface (recipe, PR)
...
Repository: localaiLicense: apache-2.0
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
dark-scarlett-v0.3-26b-a4b
Hugging Face |
GitHub |
Launch Blog |
Documentation
License: Apache 2.0 | Authors: Google DeepMind
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 small models) 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.
Featuring both Dense and Mixture-of-Experts (MoE) architectures, Gemma 4 is well-suited for tasks like text generation, coding, and reasoning. The models are available in four distinct sizes: **E2B**, **E4B**, **26B A4B**, and **31B**. Their diverse sizes make them deployable in environments ranging from high-end phones to laptops and servers, democratizing access to state-of-the-art AI.
Gemma 4 introduces key **capability and architectural advancements**:
* **Reasoning** – All models in the family are designed as highly capable reasoners, with configurable thinking modes.
...
Repository: localaiLicense: apache-2.0
privacy-filter-multilingual
A multilingual PII token-classification model: a fine-tune of
openai/privacy-filter by OpenMed. It labels every token with a BIOES tag
over 54 PII categories (217 classes) across 16 languages (ar, bn, de, en,
es, fr, hi, it, ja, ko, nl, pt, te, tr, vi, zh), spanning identity, contact,
address, financial, vehicle, digital, and crypto entities.
In LocalAI this is a PII detector for the NER redactor tier: set
known_usecases to [token_classify] (as below), and any model opts into
redaction by listing this one under pii.detectors. The detection policy
(which categories to mask vs block, and the score threshold) lives on this
model's own pii_detection block - see the overrides below. It runs locally
with no Python, served by the standalone privacy-filter backend's
TokenClassify RPC (constrained BIOES Viterbi decode into UTF-8 byte-offset
entity spans).
Architecture: gpt-oss-style sparse MoE (8 layers, 128 experts top-4, ~50M
active per token), bidirectional banded attention, o200k tokenizer; served
via the openai-privacy-filter architecture. F16, ~2.7 GB.
Repository: localaiLicense: apache-2.0
privacy-filter-nemotron
A fine-grained English PII token-classification model: a fine-tune of
openai/privacy-filter by OpenMed on NVIDIA's Nemotron-PII dataset. It labels
every token with a BIOES tag over 55 PII categories (221 classes), trading
the multilingual sibling's language breadth for category depth - identity,
contact, address, dates, government IDs, financial, healthcare, enterprise,
vehicle and digital entities (including api_key, ipv4/ipv6 and mac_address).
For multilingual text prefer privacy-filter-multilingual instead.
In LocalAI this is a PII detector for the NER redactor tier: set
known_usecases to [token_classify] (as below), and any model opts into
redaction by listing this one under pii.detectors. The detection policy
(which categories to mask vs block, and the score threshold) lives on this
model's own pii_detection block - see the overrides below. It runs locally
with no Python, served by the standalone privacy-filter backend's
TokenClassify RPC (constrained BIOES Viterbi decode into UTF-8 byte-offset
entity spans).
Architecture: gpt-oss-style sparse MoE (8 layers, d_model 640, 128 experts
top-4, ~1.5B total / ~50M active per token), bidirectional banded attention,
o200k tokenizer and a 221-way token-classification head; served via the
openai-privacy-filter architecture. F16, ~2.8 GB. (A smaller Q8_0 quant
exists on the GGUF repo for RAM-constrained use - validate it on your own
data, since for PII a single dropped span is a leak.)
Repository: localaiLicense: apache-2.0
qwopus3.5-9b-coder-mtp
# 🌟 Qwopus3.5-9B-v3.5
## 💡 Model Overview & v3.5 Design
Qwopus3.5-9B-v3.5 is a **data-scaled continuation** of the Qwopus3.5-9B-v3 model.
The training data in v3.5 is expanded to cover a broader range of domains, including mathematics, programming, puzzle-solving, multilingual dialogue, instruction-following, multi-turn interactions, and STEM-related tasks.
Qwopus3.5-9B-v3.5 is a reasoning-enhanced model based on **Qwen3.5-9B**, designed for:
- 🧩 Structured reasoning
- 🔧 Tool-augmented workflows
- 🔁 Multi-step agentic tasks
- ⚡ Token-efficient inference
Compared with Qwopus3.5-9B-v3, **3.5 version does not introduce a new architecture, RL stage, or template redesign**.
This version is trained with approximately **2× more SFT data**.
## 🎯 Motivation & Generalization Insight
The motivation behind v3.5 comes from a simple observation:
> This work is motivated by the hypothesis that scaling high-quality SFT data may further enhance the generalization ability of large language models.
In earlier Qwopus3.5 experiments, structured reasoning was observed to improve both **accuracy and efficiency**:
...
Repository: localaiLicense: apache-2.0
supergemma4-26b-uncensored-v2
Hugging Face |
GitHub |
Launch Blog |
Documentation
License: Apache 2.0 | Authors: Google DeepMind
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 small models) 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.
Featuring both Dense and Mixture-of-Experts (MoE) architectures, Gemma 4 is well-suited for tasks like text generation, coding, and reasoning. The models are available in four distinct sizes: **E2B**, **E4B**, **26B A4B**, and **31B**. Their diverse sizes make them deployable in environments ranging from high-end phones to laptops and servers, democratizing access to state-of-the-art AI.
Gemma 4 introduces key **capability and architectural advancements**:
* **Reasoning** – All models in the family are designed as highly capable reasoners, with configurable thinking modes.
...
Repository: localaiLicense: gemma
qwen_qwen3.5-2b
Qwen3.5-2B is a highly efficient, instruction-tuned multilingual language model available in various quantized GGUF formats. Optimized for llama-cpp inference, it supports chat and completion tasks with strong performance on low-RAM hardware. The model is available in multiple quantization levels ranging from Q8_0 to IQ2_M to balance quality and resource usage.
Repository: localaiLicense: apache-2.0
qwen3.5-397b-a17b
Repository: localaiLicense: apache-2.0
qwen3.5-27b
Repository: localaiLicense: apache-2.0
qwen3.5-122b-a10b
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
nanbeige4.2-3b-q8
Nanbeige4.2-3B is a compact agentic and reasoning model with 3B non-embedding parameters. Its Looped Transformer architecture reuses transformer layers to increase model capacity without adding parameters. It targets tool use, coding, mathematics, scientific reasoning, office workflows, and deep research, with English and Chinese support and a context length of up to 262K tokens. This entry uses the Q8_0 GGUF quantization.
Repository: localaiLicense: apache-2.0
nanbeige4.2-3b
Nanbeige4.2-3B is a compact agentic and reasoning model with 3B non-embedding parameters. Its Looped Transformer architecture reuses transformer layers to increase model capacity without adding parameters. It targets tool use, coding, mathematics, scientific reasoning, office workflows, and deep research, with English and Chinese support and a context length of up to 262K tokens. This entry uses the Q4_K_M GGUF quantization.
Repository: localaiLicense: apache-2.0
omnilingual-0.3b-ctc-q8-sherpa
Omnilingual ASR CTC 300M (int8) is a multilingual automatic speech recognition model supporting 1,600+ languages. Based on Meta's omniASR_CTC_300M architecture (Wav2Vec2 with CTC head), quantized to int8 for efficient inference. Uses the sherpa-onnx backend with ONNX Runtime.
Repository: localaiLicense: apache-2.0
streaming-zipformer-en-sherpa
Streaming English ASR: sherpa-onnx zipformer transducer (int8, chunk-16 left-128). Low-latency real-time transcription with endpoint detection via sherpa-onnx's online recognizer. English-only; for multilingual offline ASR see omnilingual-0.3b-ctc-q8-sherpa.
Repository: localaiLicense: apache-2.0
kokoro-multi-lang-v1.0-sherpa
Kokoro multi-lingual TTS (v1.0, int8) served through the sherpa-onnx backend with native streaming TTS. A single model covers many languages and speakers (English, Italian, Spanish, French, German and more) via a built-in voices bank; espeak-ng data and per-language lexicons ship with it. Select a speaker with the `voice` parameter (numeric speaker id) and optionally pass `language=` to hint the language.
Repository: localaiLicense: apache-2.0
supertonic-3
Supertonic multilingual text-to-speech (Supertone/supertonic-3), served through the native supertonic backend via ONNX Runtime. Lightning-fast on-device flow-matching TTS with 44.1 kHz output, 31 languages, and 10 preset voice styles (F1-F5, M1-M5). No espeak-ng dependency. Defaults to voice F1; override per request with the OpenAI `voice` field, and optionally pass `language=` (e.g. en, ko, ja, it; "na" for language-agnostic).
Repository: localaiLicense: mit
magpie-tts-cpp-357m-q8_0
Magpie TTS Multilingual 357M (C++/ggml, magpie-tts.cpp), Q8_0 (~624 MB), near-lossless
and the fastest decode. Native C++ text-to-speech for NVIDIA's Magpie TTS Multilingual
(encoder + autoregressive decoder over NanoCodec tokens) from one self-contained GGUF
(model, codec, tokenizer, G2P dictionaries). 22.05kHz mono, 5 baked voices (set `voice`
to Aria, Jason, John, Leo, Sofia or 0-4), 9+ languages (en, es, de, fr, it, pt-BR, hi,
vi, ko, plus Arabic variants). Parity-gated per component against the NeMo reference.
Repository: localaiLicense: other
Page 1