Meet Gemma 4 12B!
A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license.
Bridging the gap between edge efficiency and advanced reasoning. Here is whatโs new with Gemma 4 12B: ๐
Truly living in the future... I just set up an agent with real $$$ to trade on Robinhood using Opus 4.8 via Claude Cowork. I only deposited a small test amount of funds, will see how my trading strategy performs first!
Currently testing @MiniMax_AI M3 (newest model) with a one-shot game prompt... Also using the same prompt with Opus 4.8 to see comparative results! Game is a Halo 3 inspired FPS, single file HTML using Three.js
Hot take: we're ~6 months from nobody caring which model they use, the same way nobody cares which engine is in their Uber.
The model isn't the moat anymore, the tooling is. Codex shipping a desktop app with a built-in browser that merges coding and knowledge work? That moves the needle. A 0.1 version bump doesn't.
I also think local models catch up to cloud "close enough" sooner than the consensus expects. When that happens the model is truly interchangeable.
Codex is my favorite harness right now, and I don't see that changing... For now ๐
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.
Available today at the same price.