GigaChat3-10B-A1.8B is an efficient dialogue model in the GigaChat series. Based on the Mixture-of-Experts (MoE) architecture, it has a total of 10 billion parameters and 1.8 billion active parameters. It adopts the innovative Multi-head Latent Attention (MLA) and Multi-token Prediction (MTP) technologies, aiming to optimize inference throughput and generation speed. The model is trained on 20T tokens of diverse data and supports 10 languages including Chinese, suitable for dialogue scenarios requiring quick responses.

Natural Language ProcessingGgufMultiple Languages

VoxCPM is an innovative tokenizer-free end-to-end text-to-speech (TTS) system that overcomes the limitations of discrete tokenization by modeling speech in a continuous space. It has two core capabilities: context-aware speech generation and realistic zero-shot voice cloning. It can automatically adjust the prosody and style according to the text content and clone the speaker's timbre, accent, and emotion with just a short reference audio.

Audio ProcessingSafetensorsMultiple Languages

VibeThinker-1.5B is a 1.5-billion-parameter dense language model launched by Weibo AI. It is fine-tuned based on Qwen2.5-Math-1.5B and is specifically designed for mathematical and algorithmic coding problems. Trained using the 'Spectrum to Signal Principle' framework, it outperforms larger models in multiple math competition tests. The training cost is approximately $7,800, and it supports an output of up to about 40k tokens.

Natural Language ProcessingTransformersMultiple Languages

GigaChat3-10B-A1.8B is a dialogue model in the GigaChat series, based on the Mixture of Experts (MoE) architecture, with a total of 10 billion parameters, of which 1.8 billion are active parameters. This model uses multi-head latent attention and multi-token prediction technologies, supports a long context of 256,000 tokens, and performs excellently in multilingual dialogue and reasoning tasks.

Natural Language ProcessingSafetensorsMultiple Languages

This model is a knowledge distillation model based on the Qwen3-4B architecture, trained with approximately 54.4 million tokens generated by Gemini 2.5 Flash, aiming to integrate the behavior, inference process, and knowledge of Gemini-2.5 Flash into a single dataset.

Natural Language ProcessingTransformersEnglish

Profluent-E1 is a faithful implementation of the official Profluent E1 model, integrating Hugging Face AutoModel compatibility and focusing on the representation learning of protein sequences. It can generate high-quality protein sequence embeddings, support tasks such as masked language modeling, sequence classification, and token classification, and provides convenient embedding extraction and fine-tuning functions.

Scientific ComputingTransformers

Chonky is a multilingual paragraph segmentation model based on DistilBERT, fine-tuned on nearly 11 billion tokens from over 34 million Wikipedia articles. It can segment any natural language text into semantic chunks and supports 104 languages.

Natural Language ProcessingTransformersMultiple Languages

This model is a fine-tuned version based on meta-llama/Llama-3.3-70B-Instruct. It is trained using the mlabonne/FineTome-100k dataset, which contains 100k token data. The model is fine-tuned using the Unsloth and Huggingface TRL libraries and supports English language processing.

Natural Language ProcessingTransformersEnglish

The top model in the aquif-3.5 series, with advanced inference capabilities and a large context window of 1 million tokens, and has excellent performance in multiple benchmark tests. The comprehensive AAII score reaches 60 points.

Natural Language ProcessingTransformersMultiple Languages

The aquif-3.5 series is the masterpiece released on November 3, 2025, including two versions, Plus and Max, which provide advanced inference capabilities and an unprecedented 1 million token context window, achieving state-of-the-art performance in their respective categories.

Natural Language ProcessingTransformersMultiple Languages

Olmo 3 is a brand - new 32B parameter language model family developed by the Allen Institute for AI (Ai2), which includes variants such as Base, Instruct, and Think. This model is trained on the Dolma 3 dataset, supports long - context processing of up to 65,536 tokens, and aims to promote the development of language model science. The model is fully open - source and follows the Apache 2.0 license.

Natural Language ProcessingTransformersEnglish

WavJEPA is an audio foundation model based on the waveform joint embedding prediction architecture. It uses advanced semantic representation learning to address the deficiencies in speech unit or token-level representation learning. It significantly outperforms the state-of-the-art time-domain audio foundation models in many downstream benchmark tasks while requiring significantly less computational resources.

Audio ProcessingTransformers

Kimi Linear is an efficient hybrid linear attention architecture that outperforms traditional full attention methods in short context, long context, and reinforcement learning scenarios. It optimizes attention computation through the Kimi Delta Attention (KDA) mechanism, significantly improving performance and hardware efficiency, and is particularly good at handling long context tasks up to 1 million tokens.

Natural Language ProcessingTransformers

This is the third version of the monolingual encoder-decoder language model released by the HPLT project, specifically designed for German. It is a text-to-text transformer model trained based on the denoising objective, following the NorT5 architecture settings, and using a tokenizer trained with German HPLT data.

Natural Language ProcessingPytorchMultiple Languages

Gaperon-Young-1125-1B is a bilingual (French-English) language model with 1.5 billion parameters, developed by the ALMAnaCH team at the French National Institute for Research in Computer Science and Control (Inria Paris). The model is trained on approximately 3 trillion high-quality tokens, with a particular focus on language quality and general text generation ability rather than benchmark test optimization.

Natural Language ProcessingTransformersMultiple Languages

Gaperon-1125-24B is a bilingual (French-English) language model with 24 billion parameters. It has been trained on approximately 2 trillion tokens and can proficiently handle French, English, and coding tasks. This model uses the OLMo-2 architecture, aiming to achieve the best balance between large-scale, high-quality text generation and excellent task performance.

Natural Language ProcessingTransformersMultiple Languages

This is the GGUF quantized version of the AvitoTech/avibe model. It is based on the Qwen3-8B base model and undergoes Russian supervised fine-tuning. A custom tokenizer is used to optimize the Russian processing ability, and it is specifically designed for Russian text generation tasks.

Natural Language ProcessingGgufMultiple Languages

Svara-TTS is an open-source multilingual text-to-speech model for Indian languages, supporting 19 languages (18 Indian languages + Indian English). This model is built based on the Orpheus-style discrete audio token method, aiming to achieve clear, expressive, and low-latency speech synthesis on ordinary GPUs/CPUs.

Audio ProcessingTransformersMultiple Languages

This is a sequence-to-sequence model specifically designed to compress short user prompts (≤64 tokens) and can achieve low-latency operation of less than 100 milliseconds on modern GPUs. It serves as a lightweight preprocessing stage to support high-capacity large language models.

Natural Language ProcessingTransformers

Qwen3-Next-80B-A3B-Thinking-1M-qx64n-mlx is a large language model that adopts a mixture-of-experts architecture and Deckard mixed precision quantization technology. It has 80 billion parameters and a context length of 1M tokens. This model performs excellently in science and technology reasoning and long text processing, showing a 20 - 35% improvement in cognitive benchmark tests compared to instruction models.

Natural Language ProcessingMlx