The Wider Lens logoThe Wider Lens
Search

Model Tracker

Recently updated models from the Hugging Face Hub. Likes are lifetime figures; downloads are the last 30 days as reported by the Hub. Open-weight status is shown only when the license reliably indicates it.

AllLLMImageVideoAudioCodingEmbeddingsReasoningMultimodal

From The Wider Lens

Our own model, shown separately: it isn't part of the tracker lists below.

TheWiderLensInitiative/laya-for-clara ↗

The small model that reads every message to Clara, our private AI assistant. One model now answers five questions in about a third of a second on a CPU: chat or task, quick or deep, what kind of help it needs, whether you're following up on what she just did, and, before her browser clicks a button, whether that click needs your OK first. A fine-tune of ConvAI Innovations' Laya (ModernBERT-large, 421M parameters).

New modelapache-2.0updated 1w ago

View on Hugging FaceGet Clara free

Recently updated

Newly released

Models recently onboarded to OpenRouter, newest first (real release timestamps).

DeepSeek: DeepSeek V4.1 Flash 1017↗

DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the `deepseek-flash` API model name. V4.1 Flash is a sparse mixture-of-experts model...

via OpenRouter2d ago1.0M ctx

Elorian: Glimpse↗

Glimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than...

via OpenRouter2d ago262.1k ctx

XingChen: Xing4.0 29B A4B (free)↗

Xing4.0-29B-A4B is an open-weight mixture-of-experts language model from China Telecom's AI team (TeleAI) and the successor to the TeleChat series. It has 29B total parameters with 4B active per token...

via OpenRouter2d ago262.1k ctx

Reflection: Beam 1.0 501B A23B↗

Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per token, and a...

via OpenRouter2d ago1.0M ctx

Moondream: Moondream 3 Preview↗

Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with...

via OpenRouter3d ago32.0k ctx

Moondream: Moondream 3.1↗

Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and...

via OpenRouter3d ago32.0k ctx

Meta: Muse Spark Open↗

Muse Spark Open is a coding-optimized open-weight model. Trained specifically for agentic coding workflows, it delivers higher first-attempt accuracy, reliable tool calling, and deep codebase understanding across the repos developers...

via OpenRouter3d ago1.0M ctx

NVIDIA: Nemotron 3.5 Super VL↗

NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...

via OpenRouter3d ago1.0M ctx

NVIDIA: Nemotron 3.5 Super VL (free)↗

NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...

via OpenRouter3d ago1.0M ctx

Pareto 26.10↗

Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.

via OpenRouter3d ago1.0M ctx

Pareto Latest↗

This model always redirects to the latest model in the Pareto family.

via OpenRouter5d ago1.0M ctx

Pareto 26.9↗

Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.

via OpenRouter5d ago262.1k ctx

LiquidAI: LFM2.5-24B-A2B↗

LFM2.5-24B-A2B is a mixture-of-experts reasoning model from Liquid AI with 24B total and 2.45B active parameters. It always reasons before answering and is built for agents that call tools. Liquid...

via OpenRouter5d ago128.0k ctx

Google: Gemini 4 Argon↗

Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across...

via OpenRouter5d ago1.0M ctx

Google: Gemini 4 Argon (batch)↗

Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across...

via OpenRouter5d ago1.0M ctx

Most liked

Community favorites on the Hub.

Qwen3.8-27B↗

LLM

by Qwen

♥ 18.0k↓ 7.7M/30dAug 14, 2026open weights
transformerssafetensorsqwen3_5image-text-to-textconversational

DeepSeek-R1↗

LLM

by deepseek-ai

♥ 14.4k↓ 1.3M/30dMar 27, 2025open weights
transformerssafetensorsdeepseek_v3text-generationconversational

Kimi-K3↗

LLM

by moonshotai

♥ 11.8k↓ 1.0M/30dSep 2, 2026
transformerssafetensorskimi_k3feature-extractioncompressed-tensors

LTX-2.5↗

Image

by Lightricks

♥ 8.4k↓ 2.0M/30d1w agogated
diffusion-single-fileimage-to-videotext-to-videovideo-to-videoimage-text-to-video

JEV-27B-VL↗

LLM

by autotrust

♥ 7.7k↓ 1.8M/30d5d agoopen weights
transformerssafetensorsqwen3_5image-text-to-textjev

Kokoro-82M↗

Audio

by hexgrad

♥ 7.3k↓ 10.5M/30dApr 10, 2025open weights
text-to-speechenarxiv:2306.07691arxiv:2203.02395base_model:yl4579/StyleTTS2-LJSpeech

stable-diffusion-v1-4↗

Image

by CompVis

♥ 7.1k↓ 527.3k/30dAug 23, 2023open weights
diffuserssafetensorsstable-diffusionstable-diffusion-diffuserstext-to-image