← Back to archive

Qwen3 looks promising with open-source models and long context

Qwen3 looks promising. Their flagship Qwen3-235B-A22B, is competing head-to-head with the biggest names like DeepSeek, Grok-3, Gemini 2.5-Pro. Even their tiny models are punching way above their weight. 😳

They’re open-sourcing two MoE monsters (big + small), plus six dense models - all under Apache 2.0. Meaning.

🧠 Hybrid Thinking Modes: "fast answers" or "deep reasoning" depending on what you need.

🌎 Supports 119 languages and dialects.

🤖 Agentic capabilities + coding skills

⚡ Context lengths up to 128K, which is huge for longer conversations or documents.

You can already grab Qwen3 models on Hugging Face, ModelScope, Kaggle, and run them locally with Ollama, LMStudio, llama.cpp, etc.

Check out here: chat.qwen.ai

View on X

Qwen3 looks promising with open-source models and long context
Qwen3 looks promising with open-source models and long context

Originally posted on X

PS: I'm pretty active on 𝕏 if you'd like to follow along

Archive

2026

2025

2024

2018