Google’s latest Gemma, small yet strong for chat and generation

Qwen3 is the latest Qwen LLM, built for top-tier coding, math, reasoning, and language tasks.

Solid LLaMA 3 update, reliable for coding, chat, and Q&A tasks

OpenAI’s open-weight models designed for powerful reasoning, agentic tasks

Tiny LLM built for speed, edge devices, and local development

Gemma 4: multimodal open AI models by Google, optimized for reasoning, coding, and long context.

Distilled LLaMA by DeepSeek, fast and optimized for real-world tasks

Qwen3-Coder is Qwen’s new series of coding agent models.

The most advanced Qwen model yet, with major gains in text, vision, video, and reasoning.

Versatile Qwen update with better language skills and wider support

Google’s latest Gemma, in its QAT (quantization aware trained) variant

Efficient multimodal AI for text, image, audio, and video on low-resource devices.

Microsoft’s compact model, surprisingly capable at reasoning and code

Newest LLama 3 release with improved reasoning and generation quality

397B MoE model with 17B activation for reasoning, coding, agents, and multimodal understanding

Ministral 3: compact vision-enabled model with near-24B performance, optimized for local edge use

Efficient open model with top-tier performance and fast inference

Kimi K2 Thinking: open-source agent with deep reasoning, stable tool use, fast INT4, 256k context.

Efficient 80B MoE coding model with 3B activated params, 256K context, and agentic capabilities

DeepSeek-V3.2 boosts efficiency and reasoning with DSA, scalable RL, agentic data—IMO/IOI wins.

DeepCoder-14B-Preview is a code reasoning LLM fine-tuned to scale up to long context lengths

GLM-4.7-Flash is a top 30B-A3B MoE, balancing strong performance with efficient deployment.

SmolLM3 is a 3.1B model for efficient on-device use, with strong performance in chat

Google’s latest Gemma, small yet strong for chat and generation

Meta’s LLama 3.1: Chat-focused, benchmark-strong, multilingual-ready.

Qwen3 Embedding: multilingual models for advanced text/ranking tasks like retrieval & clustering.

Image generation model, uses a base latent diffusion model plus a refiner.

Ministral 3: compact vision-enabled model with near-24B performance, optimized for local edge use

Kimi K2 Thinking: open-source agent with deep reasoning, stable tool use, fast INT4, 256k context.

Qwen3 is the latest Qwen LLM, built for top-tier coding, math, reasoning, and language tasks.