The Frontier - SiliconNoon (Page 32)

The Frontier

DeepSeek ships V4 Pro, ending its flagship's four-month preview

DeepSeek finally took the preview label off its flagship model, and Twitch handed streamers a way out of Amazon's AI training pipeline. Two stories about the same underlying fight: who controls the models, and who controls the data. DeepSeek has released the production version of V4 Pro, its flagship model, ending a preview that ran since late April. The build — DeepSeek-V4-Pro-0813 — went live on the company's API and on OpenRouter on August 12, carrying over the preview's specs: a 1.6-trillio

The Frontier

xAI's Grok 4.6 matches GPT-5.6 Sol on intelligence index

Two frontier releases landed within hours of each other: xAI shipped Grok 4.6 with a claim of parity with OpenAI's GPT-5.6 Sol on a composite intelligence index, and DeepSeek pushed its flagship V4 Pro to general availability. A 3-billion-parameter vision model built to run on a phone rounds out the afternoon. xAI released Grok 4.6, its most capable model yet, and says it matches GPT-5.6 Sol on the Artificial Analysis Intelligence Index — a composite of nine benchmarks where both score 61, behi

The Frontier

Qwen open-sources Qwen3.8-2.4T-A95B, its first Max-class model

Wednesday's wire: Alibaba opened the weights on its most powerful model ever, Honor put a robotic arm in a phone and sold 400,000 of them before lunch, and Europe's booksellers started refusing suspicious bulk orders. Alibaba's Qwen team open-released Qwen3.8-2.4T-A95B on Wednesday — the first Qwen-Max-class model to ship as open weights, with 2.4 trillion total parameters and 95 billion active in a 512-expert mixture-of-experts architecture. The Hugging Face page went live on the exact schedul

The Frontier

Attack exposes hidden reasoning traces inside Claude, GPT, Gemini

Two stories frame the day: researchers found a way to pry open the hidden chain-of-thought inside frontier models, and Germany moved to treat Meta's AI glasses the way it treated a spy doll. Researchers from the University of Tübingen, the Max Planck Institute, MATS Research, and security firm Snyk have found a way to extract the hidden reasoning traces of frontier models — and the method turned up the strongest evidence yet that some Chinese open-weight models were trained on US rivals' privat

The Frontier

Deep Dive — Google ships the first sign-language AI in a consumer product

For the first time, a sign language is a first-class input on a phone. DeepMind's new SL2T model translates American Sign Language into English text inside Gboard and Live Transcribe on the Pixel 11 — the biggest accessibility milestone in AI this year, and a far harder engineering problem than the launch-day gloss suggests. It is also a story with open questions: whose data trained it, whose benchmark proves it, and who actually gets to use it. The milestone: sign language finally gets its di

The Frontier

Coreset pruning cuts 80% of VLM tokens, keeps accuracy intact

Vision-language models are drowning in their own pixels, and a Chinese Academy of Sciences team has a training-free fix: cut up to 80% of the visual tokens a multimodal model has to chew through — and lose nothing. The CAS Institute of Automation's Zidong Taichu team has released GMC (Grounded Message Coreset Pruning), a training-free method that cuts up to 80.2% of a vision-language model's visual tokens while retaining 97.78% of full-model capability — and in its most aggressive configuration

The Frontier

Microsoft's MAI Code 1.1 Flash trails DeepSeek on price and performance

Microsoft shipped a new coding model for GitHub Copilot that's four times cheaper than its own predecessor — and still loses to DeepSeek on both price and the benchmark that matters most. Microsoft has released MAI Code 1.1 Flash, a coding model now in production in GitHub Copilot that the company says writes better code at a quarter of the cost of the version it launched at Build in June. Microsoft claims 25 percent greater token efficiency, a 22 percent improvement on Terminal-Bench 2.1 in Co

The Frontier

MLS-Bench: frontier models can't invent new ML methods

A benchmark from UC Berkeley, Princeton and Tsinghua researchers asks whether AI systems can invent the kind of generalizable methods that built modern machine learning — residual connections, attention, Adam — or merely recombine what humans already made. The early verdict is blunt: so far, they can't, even with the answer key in hand. MLS-Bench, built by 28 researchers across UC Berkeley, Princeton, Tsinghua and seven other universities, finds that frontier models can tune machine learning sy

The Frontier

AI 101 — What is chain of thought?

Chain of thought is the step-by-step reasoning a language model generates before it gives a final answer — the "show your work" part of AI problem-solving. Instead of jumping straight from question to answer, the model writes out its intermediate steps, and that simple change measurably improves how well it handles math, logic, and anything with multiple stages. The idea surfaced in a January 2022 paper from Google Research: the authors found that asking a model to work through a problem aloud

The Frontier

Xiaomi's MiLM Plus ships PROVE, a benchmark for AI object removal

The AI-editing arms race just got a proper measuring stick: Xiaomi's research team shipped an evaluation framework that says the metrics everyone uses to judge object-removal models are out of step with human perception. Meanwhile, the open-source fine-tuning favorite Unsloth put local model training on the desktop. Xiaomi's MiLM Plus research group released PROVE, an open evaluation framework for object removal in images and video that its authors say fixes what standard metrics get wrong. The

The Frontier

Trajectory raises $40M for AI that learns without forgetting

The morning's biggest bets are on the long game: Sequoia is funding a team trying to make models keep learning after deployment, Beijing's order to unwind Meta's Manus deal is nearly done, and agent observability keeps pulling in money. Trajectory, a startup founded by former DeepMind, Apple, OpenAI, and Meta staffers, has raised $40 million led by Sequoia at a $300 million valuation to build continual learning models, according to The Information. Continual learning is the field's answer to on

The Frontier

Nvidia ships Nemotron 3.5 Lightning and an open model router

The enterprise AI story today is about choice: Nvidia shipped a fast, easily customized open model plus a router that picks between models, IBM bet $240 million on open-model inference, and Spotify decided AI-generated artist identities shouldn't get recommendations. Nvidia released Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts model aimed at the high-volume, always-on workloads that power AI agents, alongside NeMo Switchyard, an open-source library that routes each step of

See all