The Frontier - SiliconNoon (Page 23)

The Frontier

DeepSeek set to raise $7.4B at a $74B valuation

The biggest private model lab in China is about to get materially bigger. DeepSeek is set to raise $7.4 billion from existing and new investors at a $74 billion valuation, according to the Wall Street Journal — the capital earmarked for research and for an aggressive expansion into the computing infrastructure it currently rents. The deal would cement DeepSeek not just as an open-weights disruptor, but as a company with the balance sheet to build the scale behind its famous price advantage. The

The Frontier

Zhipu's mystery 'Niulai' model rockets to No.2 in global usage in four days

A mystery Chinese model that appeared seemingly out of nowhere has become one of the most-used AI models in the world within days, and the rankings it just cracked say something broader about where frontier computing is heading. A mystery Chinese model called "Niulai" has jumped to No. 2 in global AI usage just four days after appearing, while DeepSeek's V4-Flash holds the top spot for a third consecutive week. Reports attribute the model to Beijing's Zhipu, which has reportedly trained it to r

The Frontier

Deep Cogito raises $43M to make AI that improves itself

A small lab founded by two ex-Google engineers is betting that the next leap in AI intelligence won't come from bigger pre-training runs — it'll come from teaching models to learn from their own reasoning. Deep Cogito has closed a $43 million Series A to build what it calls the "post-training engine" for frontier models, and the pitch is quickly becoming one of the industry's core debates. Deep Cogito raised a $43 million Series A to build a "post-training engine" that makes AI models progressi

The Frontier

Google debuts Gemini 3.5 Transcribe, a smarter speech-to-text model

Two Google model releases dominate the AI news cycle this afternoon — neither a weight-dump spectacle, both quietly aimed at making AI's messier inputs tractable. One turns raw audio into clean, structured text; the other turns a stream of glucose readings into transferable medical insight. Here's the rundown. Google opens Gemini 3.5 Transcribe to developers, a speech-to-text model built to clean up messy audio. Announced Wednesday in public preview for the Gemini API and the Gemini Enterprise

The Frontier

Alibaba's Qwen3.8-Flash-Next hits Opus-class at one-ninth cost

The cost-efficiency race moved another notch on Tuesday night: Alibaba open-sourced Qwen3.8-Flash-Next, a sparse mixture-of-experts model that beats much bigger and pricier rivals on coding and office benchmarks while training for roughly a ninth of the bill. It lands hours after the market absorbed two more cheap models in the same vein, a sign that frontier-grade performance is increasingly a pricing story. Alibaba's Qwen team released and open-sourced Qwen3.8-Flash-Next, a model that packs 1

The Frontier

GLM-5.3-Flash proves frontier intelligence is now a Chinese-chip economics story

The model that quietly ruled the OpenRouter leaderboards all week finally acknowledged its identity today, and the technical detail buried in the announcement is more consequential than the reveal itself. Z.ai confirmed that GLM-5.3-Flash — launched this morning — is the same engine that ran anonymously as "ox-alpha" and became the week's most-used open model. That much we expected, having watched sleuths chase the attribution from Mystery model Ox Alpha tops GPT-5.6 — sleuths say it's Zhipu's G

The Frontier

Z.ai's GLM-5.3-Flash tops benchmarks at one-tenth the price

The mystery model that ruled the leaderboards all week finally showed its face today — and the reveal came with a price tag that's reshaping the frontier. Z.ai has confirmed that GLM-5.3-Flash — the model it released today — is the same engine that topped benchmarks anonymously as "Ox Alpha" this week. The first natively multimodal model in the GLM-5 series packs 320 billion total parameters with just 18 billion active, a sparse MoE design the company says lets it beat its predecessor GLM-5.2 a

The Frontier

Qwen3.8-27B ships a fully NVFP4 build that keeps its benchmark scores at a third of the size

A quantization team called QUASAR-QAT has released a fully 4-bit build of Qwen3.8-27B that keeps nearly all of the original model's accuracy — and it did it by retraining the quantization in, not by rounding weights after the fact. A fully NVFP4-quantized checkpoint of Qwen3.8-27B is out, and it posts scores within a rounding error of the unquantized original while cutting the download from roughly 56 GB to under 20 GB. The release comes from a group publishing under the name QUASAR-QAT, and th

The Frontier

IBM's open Granite 4.2 models learn agent skills via reinforcement learning

A string of open-weight releases this week keeps pushing what small, deployable models can do. IBM is the latest, rolling out a reasoning-tuned Granite family that trains the bigger variants to actually use tools and code in real sandboxes before shipping them under a permissive Apache 2.0 license. IBM released Granite 4.2, a family of dense, open-weight reasoning models in 3B, 8B, and 30B sizes, its first to be post-trained with reinforcement learning that teaches them agent-style behavior. Ea

The Frontier

Apple's STARFlow2 makes normalizing flows a rival to diffusion

While most of the industry keeps refining diffusion, Apple's research group is betting on a cleaner architecture for models that understand and generate text and images in one system. Apple's machine learning research group published STARFlow2, a unified multimodal model built on autoregressive normalizing flows — a direct challenge to the diffusion pipelines that power nearly every image generator today. The paper's starting observation is disarmingly simple: an autoregressive normalizing flow

The Frontier

Accelerated Understanding launches a physics foundation model

AI's frontier-pushing crowd spent the week cleaving to language and agentic models — but one of the field's better-known researchers used Tuesday to bet on a different kind of foundation model: one that predicts how the physical world changes, not how the world talks. Caltech professor Anima Anandkumar, a former Nvidia machine-learning research director, unveiled a startup and a model aimed squarely at enterprise engineering, promising a single system for chip design, weather, and geology. Acce

The Frontier

Claude Opus 5 lifts its test pass rate to 88.6% — and writes 2.3x more code

Code-analysis firm Sonar put Anthropic's Claude Opus 5 through its standard LLM evaluation and found the model passes more of the tests it's given — hitting 88.6% on executable coding tasks versus 82.9% for the Opus 4.8 it replaces — while generating more than twice as much code along the way. The pass-rate gain is the headline. Across 544 HumanEval and MBPP tasks with executable tests, Opus 5 landed at 88.6%, a 5.7-point step up from Opus 4.8's 82.9%. The three Opus releases before it all clus

See all