Kimi K2.7-Code cuts thinking tokens 30% — but practitioners say the benchmarks don't check out

Kimi K2.7-Code claims 30% fewer thinking tokens and a drop-in API swap path, but independent benchmarks show kernel ...

Google's DiffusionGemma generates 256 tokens in parallel and self-corrects as it goes

Google's open-source diffusion language model generates 256 tokens in parallel and self-corrects, hitting 4x speed on one GPU ...

NanoClaw and JFrog launch 'immune system' to block AI agents from downloading malicious code

This ensures that all agent activity adheres to the company’s specific commercial licenses, internal security policies, ...

Microsoft’s open-source SkillOpt automatically upgrades AI agent skills without touching model weights

Microsoft's SkillOpt brings deep-learning discipline to AI agent skills, replacing manual prompt tweaking with mathematically ...

PixelRAG beats text parsers on accuracy and cuts AI agent token costs 10x

UC Berkeley's PixelRAG renders pages as screenshots instead of parsing text, boosting RAG accuracy by up to 18.1% and cutting AI agent token costs 10x.

Researchers say they trained a foundation model from scratch for about $1,500

Sapient researchers trained a 1B reasoning model on just 40B tokens — scoring competitively with 2B-7B models at a fraction ...

Why AI that works in the lab often fails in production — and what actually fixes it

As AI continues to evolve, leaders should invest not only in tools, but also in R&D processes and cultural foundations that ...

Surprise upset: GPT-5.5 beats Claude Fable 5 on brutal new Agents’ Last Exam benchmark

The victory of GPT-5.5 aligns with recent third-party analysis suggesting that OpenAI's models are currently superior at ...

Context compression finally works in production: new research cuts LLM input 16x without the accuracy hit

LCLMs compress LLM context before decode — 8.8x faster at 16x compression, beating every KV cache method tested. Open-sourced by NYU and Columbia.

What AI benchmarks miss about real-world performance

Enterprise AI teams have spent years solving for compute, securing GPU allocations, negotiating cloud capacity, and ...

Apple’s new Siri AI is more than just a smarter assistant — it's a new enterprise app layer

For enterprise technology leaders, it means Apple’s devices could soon include a native AI assistant that can act across ...

Anthropic brings Mythos to the masses with Claude Fable 5, its most powerful generally available model ever

Anthropic is pricing both Fable 5 and Mythos 5 at $10 per million input tokens and $50 per million output tokens. The company ...