Another day, another AI model from Google. This time, Google DeepMind has released a new member of the Gemma 4 open model family, but it’s fundamentally different from the rest of the lineup.

DSpark can make decoding faster, but acceptance quality still determines how much speed the system actually realizes.

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

Rather than generating text word by word, Google's experimental open-source model drafts entire passages simultaneously using diffusion, resulting in up to 4x faster inference.

Venice AI has raised a $65 million Series A round led by Dragonfly Capital, giving the privacy-focused AI company a $1 billion valuation roughly two years after its public launch. The Las Vegas-based ...

NVIDIA diffusion language model Nemotron TwoTower achieves 2.42x LLM inference throughput without a full retraining run, ...

Texas passed what may be the first state-mandated book list for public school students. It focuses on classic literature and ...

Nano Banana 2 Lite is the fastest and cheapest model in the family at four seconds and $0.034 per thousand, while Gemini Omni ...

Seedance 2.0 is ByteDance's flagship video generation model, released in 2026. It produces cinematic video up to 1080p natively, with synchronized audio, accurate lip-sync, and 4K available through ...

The time has come for the data center industry to agree upon a new efficiency metric, one that is both intuitive and ...

As the US celebrates 250 years, the Declaration of Independence has been curiously absent. Yet its language on the consent of the governed is more relevant than ever ...

Mistral OCR 4 brings bounding boxes, typed-block classification, and 170-language document extraction to enterprises that ...