Embodied AI2026
Predicts eight future [x,y,yaw] waypoints for language-conditioned embodied person-following using fused DINOv3 and SigLIP visual features; trained with quality-driven, DAgger-style self-evolving data and optimized for on-device inference (~5+ FPS, ~180 ms).
AI Dataset2026
Provides verified, model-attested end-to-end agent coding and debugging trajectories (JSONL). Each whole-session trace was produced by moonshotai/kimi-k3 on the pi/openrouter runtime, passed acceptance tests and independent model screening — useful for SFT, distillation, and analyzing tool-use behavior.
AI Dataset2026
Provides intermediate pretraining checkpoints for the Aether-7B-5Attn base model to enable reproducible training-dynamics research. Includes three raw checkpoints (110k, 115k, 162k steps) packaged with model.safetensors, config, and tokenizer; uses a custom aether_v2_7way architecture requiring the aether_pkg loader.
AI Dataset2026
Provides newline-delimited JSON agent session traces (5 files) generated with Teich for moonshotai/kimi-k3, including recovered and embedded tool-schema snapshots so traces remain training-ready even when tools weren't invoked; includes guidance for Teich data preparation and conversion.
AI Model2026
GGUF-quantized releases (NEO IMATRIX + MTP) of a multi-stage fine-tuned, uncensored Qwen3.5-9B model with vision enabled and a native 256k context window—optimized for instruction following, reasoning and image-text-to-text workflows; released under Apache-2.0.
AI Dataset2026
Provides 2,000 hours of synchronized, high‑fidelity robot‑free bimanual manipulation demonstrations with multi‑view video, calibrated end‑effector trajectories, gripper states, and language annotations. Curated from a 20,000+ hour corpus; features 6 camera views, ~3 mm pose accuracy, <40 µs cross‑sensor sync, and LeRobot v3‑style Parquet+MP4 export under CC BY 4.0.
AI Dataset2026
Provides under-1K JSON agent-trace records documenting model refusal responses and forensic metadata — useful for evaluating refusal-detection, audit pipelines, and safety analysis; small size limits large-scale statistical studies.
AI Model2026
Adds vision to GLM-5.2 by attaching a MoonViT encoder and a trained 49.5M-parameter PatchMerger projector to enable image→text multimodal reasoning; text and vision backbones are frozen, uses NVFP4 quantized weights and targets Blackwell B200 GPUs.
AI Model2026
Generates and reasons about multimodal physical-world content—text, images, video and action trajectories—conditioned on text, images, video and robot/vehicle action inputs. An edge-sized (4B) Mixture‑of‑Transformers omni-model optimized for single‑GPU inference and Physical AI tasks (image→video, action prediction, robot policy).
AI Dataset2026
A multiple-choice benchmark for evaluating language-model arithmetic: 1,000 continuation-style elementary word problems (4 choices, balanced labels) organized by topic, grade band, and difficulty. Designed for base-model continuation log-likelihood scoring; released under Apache-2.0.
AI Model2026
Efficient 4B native-resolution diffusion foundation model for text-to-image generation and instruction-based image editing. Uses a lightweight Mage‑VAE tokenizer and a 4B NR‑MMDiT backbone to produce 512–2048 outputs with low memory and fast inference; ships in Base, RL-aligned and few-step Turbo variants.
AI Model2026
Provides GGUF-format quantized shards of Laguna S 2.1 for local or self-hosted inference—packaged for llama.cpp/llama-server and usable with vLLM/Transformers runtimes; targeted at long-context, agentic coding workloads.