Alibaba Enters the AI Video Race with Wan 2.5

• The launch of Alibaba's Wan 2.5 marks a significant development in the competitive AI video market. This article provides an in-depth analysis of Wan 2.5's capabilities, its performance on real-world examples, and what its entry means for the future of generative AI tools.

10 Best AI Video Tools for Marketing Teams in 2026

• AI ad production has split into distinct categories: URL-to-ad tools, avatar platforms, content repurposing tools, and cinematic generators. Here is what each one actually does and where it falls short.

Apple's new desktop computers are designed specifically for local AI development

Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original

DeepSeek's new vision model nears Opus-level multimodal performance at Flash pricing, while a fresh study shows CLI agents beat MCP by up to 28x on cost.

• DeepSeek shipped V4-Flash-Vision-Exp, adding image/screenshot/document understanding to V4-Flash at no price premium, approaching Opus-4.8 on multimodal agent benchmarks (384 tokens/image cap, free Files API).
• A study across 7 agent scaffolds found CLI-based agents are 5-28x cheaper than MCP-based agents for the same tasks.
• Claude Code Remote Control now supports auto-reconnect, starting sessions from phone to any linked machine, and working slash commands (/clear, /compact, /diff) on mobile.
• Anthropic opened Claude Security scans (now powered by its top model, Mythos 5) to all Enterprise customers and launched a $35M Defender Advantage Fund for open-source vulnerability patching.
• Anthropic internally found its own interpretability tools add zero benefit over simply reading model transcripts directly.
• NVIDIA's coding agent scored 100% on ARC-AGI-3 with zero instructions or stated goals.

Anthropic polishes Claude's UX infrastructure (renderer + enterprise auth) while Pi's disk-spill trick slashes agent context costs by up to 88%.

• Anthropic rebuilt Claude's streaming renderer to only update actively-changing text instead of re-rendering full replies, cutting worst-case freezes 4.5x and holding steady 120fps on MacBooks.
• Anthropic shipped enterprise-managed auth for MCP connectors, removing manual OAuth setup for organizations connecting Claude to internal tools.
• Pi's coding agent uses a 'prune + spill' memory pattern — full tool output is saved to disk while context holds only a file path — cutting context usage 26-35% and processing costs up to 88% across 19 real sessions.
• An open-source Claude-based tool (built by a laid-off geophysicist) uses a drafter-reviewer agent pattern to auto-write tailored CVs and cover letters; landed 20 interviews from 69 applications.
• VoroTracing renders 3D scenes at 623 FPS, beating Gaussian Splatting by 2.8x, signaling a real-time rendering stack upgrade.
• Thinking Machines launched $50,000 in grants for open-weight AI safety research.

OpenAI's price war and Thomson Reuters' proprietary LLM show enterprises are getting disciplined about what AI intelligence is actually worth.

• OpenAI cut GPT-5.6 Sol prices over 20% (now $4/$20 per million input/output tokens), undercutting Anthropic's Fable 5 ($10/$50) and even Opus 5 ($5/$25).
• Ramp payments data shows enterprises are skipping top-tier models: Fable 5 is only 6% of Anthropic token spend, while cheaper Opus 5 has overtaken it in corporate usage.
• Thomson Reuters launched 'Thomson,' a proprietary LLM built on an open-source base plus decades of Westlaw/Reuters/Practical Law data, for $40M—far below frontier training costs—now live in CoCounsel Legal.
• Indeed's new AI exposure index ranks San Jose (59), Seattle (57), Washington DC (54), San Francisco (53), and Austin (52) as most AI-exposed metros, driven by software/knowledge work; trade and hands-on jobs remain resilient.
• Nvidia is putting $6B into a US open-source AI ecosystem and reportedly negotiating a Perplexity investment at a $30B valuation.

Wire It, Run It, Deploy It: AI Workflows in Gradio

An anonymous 'stealth model' called Ox Alpha is beating GPT-5.6 and Claude on coding benchmarks, while a tiny Nvidia Nemotron-based model is outperforming frontier APIs on a real sales task at 100x lower cost.

• Ox Alpha, a stealth model on OpenRouter with a 1.05M-token context window, reportedly beats Anthropic's Fable 5 and OpenAI's GPT-5.6 Sol on Deep SWE coding evals; origin unconfirmed (speculation points to Google Gemini, Microsoft MAI, or Z.ai's GLM)
• NextLM's Savant 3.5, built on ~30B-parameter Nvidia Nemotron models, beat GPT-5.6 Sol, Grok 4.5, Fable 5 and Opus 5 at identifying sales prospects — at $0.003-0.011 per 1,000 scores vs $0.26-5.11 for frontier APIs
• OpenAI cut GPT-5.6 Sol API pricing by over 20%, signaling continued price compression at the frontier
• DeepSeek unveiled an experimental multimodal model positioned to rival Anthropic's Opus 4.8
• Anthropic is reportedly exploring a $100B IPO and just hired Google's ex-TPU chief Amir Salek for an in-house chip push
• Hugging Face is reportedly in M&A talks at a $13B valuation

Granite 4.2 LLMs: How They're Built

llm-anthropic 0.27

• <p><strong>Release:</strong> <a href="https://github.com/simonw/llm-anthropic/releases/tag/0.27">llm-anthropic 0.27</a></p> <p>This release of the Anthropic plugin for <a href="https://llm.datasette.io/">LLM</a> mainly provides compatibility with the recently released <a href="https://github.com/ant