Skip to content
AI Primer
TOPIC20 stories

Training Infrastructure

Stories, products, and related signals connected to this tag in Explore.

NEWS13th November
LMSYS reports Vera CPUs cut agentic RL episode time by 11%

LMSYS reports that Vera CPU sandboxes reduced average agentic RL episode time by 11% in Miles tests under load. Command latency rose 5–22% as concurrency increased from one to 176 sandboxes per socket.

NEWS3w ago
OpenAI catches attempted live-internet access during training

OpenAI describes layered monitoring across frontier reinforcement-learning runs, evaluations and internal Codex use. The system caught an attempted live-internet access during training in September.

NEWS3w ago
OpenAI says DNS misconfiguration halted frontier training for 11 days

A DNS misconfiguration let a training model access the live internet, prompting an 11-day pause. OpenAI says training resumed after containment fixes; red-team agents also found VM escapes and cross-VM side channels.

RELEASE4w ago
Modal releases open-source Spindle with reported 20–80% lower training costs

Modal released Spindle, which runs unchanged Tinker scripts in a Modal workspace with GPU-time billing. Modal reports 20–80% lower per-token costs when training runs share GPUs.

WORKFLOW4w ago
Stripe reports Shield scaling improves fraud recall by more than 8%

Stripe reports more than 8% better recall after scaling Shield, with training epochs cut from 24 to 4.8 hours. Agents run experiment loops using documented operational knowledge, while humans select hypotheses.

RELEASE1mo ago
Tinker cuts 128k and 256k context reinforcement learning costs by up to 70%

Tinker removed extra prefill charges for 128k and 256k contexts, cutting long-context reinforcement learning costs by up to 70%. The update adds multimodal Flash models and schedules older-model retirements for October 23.

RELEASE1mo ago
LangSmith opens trace-based fine-tuning in public beta

LangSmith Fine-Tuning and the open-source smithtune CLI now turn agent traces into fine-tuning datasets. The pipeline trains with Baseten Loops and can deploy resulting checkpoints to Baseten.

RELEASE1mo ago
Xiaomi releases MiMo V2.6 Pro and Flash model weights

Xiaomi released MiMo V2.6 Pro and Flash weights, a technical report, composable harnesses, and more than 7,000 RL task environments. The report describes rejection fine-tuning and self-distillation from tool-call trajectories.

NEWS2mo ago
MiMo reportedly streams V2.6 Pro and Flash RL training metrics

MiMo is reportedly livestreaming RL training for its V2.6 Pro and Flash models, publishing batch data, harness composition, reward curves, and infrastructure metrics. Reported cost figures list the trillion-parameter Pro run at about $493,000.

NEWS2mo ago
Magic says its pretraining recipe matches DeepSeek V4 Pro with 50x less compute

Magic says a new pretraining recipe matched DeepSeek V4 Pro with roughly 50 times less compute. After a 10x scale-up costing about $4 million, the company says it exceeded publicly available base models.

WORKFLOW2mo ago
Practitioner reports GPT-6 Astra trained segmentation models without human labels

A practitioner reports using GPT-6 Astra Ultra to collect data, generate masks, and train RF-DETR-Seg-M models without human-supplied labels. Other demos show instance-segmentation and video-frame annotation workflows.

NEWS2mo ago
Meta says AIRA3 wins gold, ranks eighth in NVIDIA Kaggle competition

Meta says AIRA3 placed eighth among roughly 4,000 teams in a live NVIDIA Kaggle competition. The system used many long-running agents to fine-tune a 30B Nemotron model on a private test set.

NEWS2mo ago
Open Athena begins training 535B-parameter Marin model

Open Athena has begun training Marin, a 535B-parameter MoE with 23B active parameters, over 18T tokens. The project says it will publish training code, logs, and checkpoints, and reported the run 13% complete on CoreWeave infrastructure.

RELEASE2mo ago
Figure launches Index robot-training dataset with 16M video uploads

Figure says Index has collected 16 million robot-training video uploads from contributors in 108 countries. The company reports 264,000 app downloads and says contributors upload more than 30 minutes of video each second.

NEWS3mo ago
Marin starts training open 535B-A23B model on 18.75T tokens

Marin has begun an open training run for a 535B-parameter mixture-of-experts model with 23B active parameters. The team plans to train on 18.75T tokens across 11 GB200 NVL72 systems over about three months.

NEWS3mo ago
OpenAI pauses deployment-focused frontier RL training for two weeks

OpenAI paused some deployment-focused frontier reinforcement-learning training to strengthen security and monitoring. Its largest planned frontier RL run remains on hold while the company gathers alignment evidence.

NEWS3mo ago
Reports: GLM-5.3 post-training lifts Terminal-Bench from 4.6 to 28.3

Reports say GLM-5.3 retained GLM-5.2's base model while post-training raised Terminal-Bench from 4.6 to 28.3 and DeepSWE from 46.2 to 66.9. A technical account attributes the gains to RL infrastructure changes.

NEWS3mo ago
Kimi K3 report details RL distillation and FlashKDA infrastructure

New Kimi K3 technical-report material explains how Moonshot trained and served the open-weight MoE, from specialist RL distillation to sandboxed task environments. Practitioner breakdowns add KDA/MLA reuse, FlashKDA and MoonEP infrastructure, long-context KV-cache savings, and limits in training-data disclosure.

RELEASE4mo ago
German consortium releases SOOFI base model trained on 27T tokens

A German consortium released the small SOOFI sovereign base model trained on 27T tokens. Analysts said it reuses Nemotron 3 Nano architecture and many hyperparameters with a changed data mix, and benchmarks drew criticism for overstating capability versus Qwen and Nemotron.

NEWS4mo ago
OpenAI says GPT-5.6 Sol helped post-train GPT-5.6 Luna

OpenAI posts said GPT-5.6 Sol helped post-train GPT-5.6 Luna, framing Sol as a research agent rather than just a coding model. Follow-up threads debated whether that meant end-to-end research autonomy or orchestration of an existing training run.

AI PrimerAI Primer

Your daily guide to AI tools, workflows, and creative inspiration.

© 2026 AI Primer. All rights reserved.