Arcee AI Trinity-Large-Thinking

Arcee AI Trinity-Large-Thinking is a 398B Apache 2.0 sparse MoE with 13B active parameters, chain-of-thought reasoning, 512k context, and PinchBench scores just behind Claude Opus 4.6.

Meta Muse Spark

Meta launches Muse Spark (codename Avocado), its first proprietary model from Superintelligence Labs. Open-weight versions planned. Mango world model for video still in development.

AI2 OLMo 3.1

The Allen Institute for AI (AI2) has released OLMo 3.1, a family of fully open 32 billion parameter reasoning models. Unlike most open-weight releases, OLMo 3.1 ships with the complete training data, training code, evaluation scripts, and intermediate checkpoints, making it the most transparent frontier model ever published. The Think 32B variant gains 5+ points on AIME, 4+ on ZebraLogic, 4+ on IFEval, and over 20 points on IFBench compared to OLMo 3. AI2 calls the Instruct 32B variant the most capable fully open chat model to date.