Trending

Gemma3 – The current strongest model that fits on a single GPU

Sidekick: Local-first native macOS LLM app

A Practical Guide to Running Local LLMs

The Einstein AI Model

People are just as bad as my LLMs

Asking LLMs to create my game Shepard's Dog

Meta must defend claim it stripped copyright info from Llama's training fodder

I've been using Claude Code for a couple of days

From Prompt to Adventures:Creating Games with LLMs and Restate Durable Functions

qwq and gemma-3 added to long context benchmark

I deleted all my previous models after using (Reka flash 3 , 21B model) this one deserve more attention, tested it in coding and its so good

First attempt at AI agent. Where do I start

HowTo: Decentralized LLM on Akash, IPFS & Pocket Network, could this run LLaMA?

KoboldCPP 1.86 just dropped with support of Gemma-3

Microsoft

Save item

Phi-4-mini-instruct

Task: text-generation

Bartowski

Save item

Qwen_QwQ-32B-GGUF

Task: text-generation

GSAI-ML

Save item

LLaDA-8B-Instruct

Task: text-generation

AI21

Save item

AI21-Jamba-Large-1.6

Task: text-generation

Qwen

Save item

QwQ-32B-AWQ

Task: text-generation

Qwen

Save item

QwQ-32B-Preview

Task: text-generation

Unsloth AI

Save item

QwQ-32B-GGUF

Task: text-generation

Meta Llama

Save item

Llama-3.3-70B-Instruct

Task: text-generation

run-llama

Save item

llama_index

LlamaIndex is the leading framework for building LLM-powered agents over your data.

langchain-ai

Save item

langchain

🦜🔗 Build context-aware reasoning applications

BlinkDL

Save item

RWKV-LM

RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RNN and transformer - great performance, linear time, constant space (no kv-cache), fast training, infinite ctx_len, and free sentence embedding.