Grok 4.6

SpaceXAI's new proprietary frontier model for coding, long-running agents, knowledge work, and ambitious interactive and visual tasks. It extends Grok 4.5 with a longer supplemental-training run, regenerated SFT trajectories, and agentic reinforcement learning. Supports text and image input, 500K context, and low through xhigh reasoning effort.

SpaceXAI reports an AA Intelligence Index of 61, CursorBench 3.2 at 69.9%, DeepSWE 1.1 at 65.9%, and APEX-Agents at 57.5%. API pricing starts at $2/M input and $6/M output tokens below 200K prompt length. Parameter count and training scale remain undisclosed.