Despite having only 35B parameters, it even surpasses Qwen 3.5-397B on Terminal-Bench 2.1 (64.4 vs. 53.5)

At 770M parameters, a looped model achieves the downstream quality of a 1.3B fixed-depth Transformer trained on the same data — roughly half the parameters for the same quality.

we can reach the same capabilities with over an order of magnitude less compute than our previous model, Llama 4 Maverick.

This observation challenges widely accepted test-time scaling laws, leading us to hypothesize that errors within the reasoning path scale concurrently with test time.