Provides a large corpus of runnable 3D programs and a multi-agent harness that generates, executes, measures, visually judges, and repairs executable 3D code to produce editable, verifiable 3D scenes. Includes ~269K runnable programs, 2.03M render–caption pairs, and an automated execution+repair pipeline.
Answers spatial choice, scalar, and point questions by classifying among discretized candidates using pretrained attention projections from a Qwen3.5-4B backbone instead of autoregressive text generation. Trains on converted spatial QA data, adopts coarse-to-fine refinement and joint multi-stage supervision, and reports higher accuracy with several-fold faster inference than autoregressive baselines.
AI Model2026
Provides open-weight, natively multimodal LLM optimized for long-horizon multi-step reasoning and coding/agentic workflows; 816B total / 42B active parameters, 1M-token context window, available as BF16 and NVFP4 checkpoints under the Muse Model License.