What changed in AI today

• Mixture-of-experts model with about 816B total parameters, roughly 42B active per token.
• Accepts text, images, audio and video; context window reaches up to one million tokens.
• Meta says it was trained for agentic coding, including multi-file refactors and debugging.
• Two checkpoints are listed: a BF16 version and an NVFP4 version aimed at NVIDIA Blackwell GPUs.

Why it matters for marketers: Content and developer teams can self-host a long-context multimodal model and test it on screenshots, voice notes and screen recordings alongside code.

• 24B total parameters with 2B active per token, built as a reasoning model that thinks before answering.
• Scores 48.09 on LiveClaw versus 38.19 for Qwen3.6-35B-A3B, a model with roughly 50% more parameters.
• Decodes 184 tokens/s on an Apple M5 Max and 114 tokens/s on a Ryzen AI Max+ 395, per the source.
• Open-weight and available on Hugging Face now; supports llama.cpp, MLX, vLLM, SGLang, and ONNX from day one.

Why it matters for marketers: Content teams building tool-using assistants can run a local agent model, but should pair it with retrieval or search for knowledge-heavy tasks.

• Ai2's AI Infrastructure team swapped a priority-based scheduler for GPU time budgets, hierarchical fair-share allocation, and a time-slicing contract.
• Clusters of 88 to 1024 GPUs serve about 150 internal researchers, with outstanding requests reportedly 2-3x available GPUs.
• Workloads must declare a minimum runtime; unbudgeted, zero-runtime jobs are free but always preemptible.
• Automated draining of unhealthy hosts reportedly cut repairs needing a human-in-the-loop by 74%.

Why it matters for marketers: Teams sharing GPU clusters can see how budget-based allocation and minimum runtime contracts reduce idle capacity and on-call toil.

• A 0.8B prompt rewriter costs about USD 16 total, runs on a CPU and returns valid output 99.7% of the time.
• Citrus disease VLM fine-tune raised correct diagnosis on 335 test photos from 14.9% to 52.8%, for about USD 1.90.
• Agate 4-step distillation cut generation from 50 steps to 4; GenEval rose from 0.509 to 0.536 after a second run.
• Author says a prompt should include a baseline run, a smoke test and a spending cap, and that the agent starts with zero budget.

Why it matters for marketers: Content teams can use the author's prompt structure, including baselines, smoke tests and budget caps, to scope small custom model projects before committing compute.