
MBZUAI Opens RL360, the Agent Training Engine Behind K2 Horizon
IFM open-sources the agentic reinforcement learning stack behind K2 Horizon, running 8,192 concurrent sandboxes and 1.3M rollouts across seven days.
A focused feed of models, agents, research and open-source releases for people building with AI.

IFM open-sources the agentic reinforcement learning stack behind K2 Horizon, running 8,192 concurrent sandboxes and 1.3M rollouts across seven days.

New research shows hidden model preferences slip through distillation even when teacher and student share no base model, data, or architecture.

Prime Intellect ships a new default RL loss built for async training, plus a 1.15x throughput bump over torchtitan on Qwen3-235B.

Meta's Prometheus supercluster hits 85% goodput and 47K GPU-hour mean time between interruptions across hundreds of thousands of GPUs training Muse Wonder.

A new analysis settles whether momentum-based optimizers can get stuck at saddle points, and designs a variant that escapes faster without hurting convergence.

A one-shot optimal transport method aggregates independently trained mixture-of-experts models into a single global estimator without pooling raw data.

A new unified framework reframes VAE design for diffusion as a spectrum-matching problem, boosting generation quality on CelebA and ImageNet.

Zephon plugs into TorchTitan to deliver identical global batches across GPU counts, so a job resumed on different hardware trains on the exact same data.

A new study shows that proxy models used to pick pretraining data can flip their verdicts with tiny learning rate changes, with a simple fix.

Alibaba's Qwen team unveils an elastic RL framework that trains agents on hours-long tasks, cutting training time by up to 44% while lifting repo-generation accuracy by 6 points.

New research finds smaller batches and tuned recipes dramatically improve pretraining when you have more compute than unique data.

Ai2 replaced its priority-based GPU scheduler with time budgets and fair-share allocation, cutting median queue wait from 5 minutes to 24 seconds.

A new residual connection design borrows the multi-head trick from attention and applies it along network depth, delivering 1.58x compute speedup over standard residuals at MoE scale.

Success-Guided Sampling fixes a wasted-batch problem in massively parallel RL, letting PPO scale past a million simulated robots and transfer to real hardware.

A new paper fits a scaling law to robotic world models, trains an 8B JEPA on 15,000 hours of robot video, and plans on real hardware.

Zyphra researchers cut MoE all-to-all communication by up to 2.63x on AMD MI300X GPUs by exploiting predictable token routing patterns, with no changes to the model.

A new study finds no principled hyperparameter scaling rule consistently preserves optimizer performance across batch sizes, undermining single-batch benchmarks.

Stanford researchers show that steering RL exploration along human hand coordination patterns, rather than wiggling joints independently, dramatically speeds up dexterous skill learning.

A new RL post-training method shrinks memory by 45.7% using low-rank gradient sketches, letting a 27B model train on a single eight-GPU node.

A new paper argues the right dropout placement, not just positional encoding, lets transformers extrapolate up to 64x past their training context length.

A new survey by Abdelkader Baggag and Yousef Saad bridges the gap between classical numerical linear algebra and the matrix-heavy machinery powering modern LLMs.

Google's Project Suncatcher just put four Trillium TPUs into low Earth orbit on a Planet-built satellite, testing whether AI compute can survive and run in space.

Google rebuilt Federated Learning around Trusted Execution Environments, moving gradient computation server-side with externally verifiable differential privacy guarantees.

A new world action model skips latent action detours and learns robot control priors straight from raw video, posting 73.6% on RoboTwin Random.
No stories match these filters.