MarkTechPost (5 days ago)
JetBrains released Mellum2.1, an Apache 2.0, 12B mixture-of-experts thinking model with 2.5B active parameters. RL in real repositories lifted its SWE-bench Verified score from 2.0 to 47.0.
Read the full story on MarkTechPost.
Alongside the new Apple HomeView, upgraded Apple TV 4K, and Apple Home Collection, the company also unveiled the HomePod mini 2. The updated smart speaker brings Siri AI, a new Sound Glow feature, support for pairing up to 4 speakers, and new color options.
Flox’s edge devices can identify animals and use species-specific sounds to drive them away from critical infrastructure.
Google has released Gemini 4 Argon globally to paying users, in the Gemini app, in Search and through the API. The $2 and $10 per million token prices are a launch discount against standard rates of $4 and $20. Google has released Gemini 4 Argon to anyone willing to pay for it.
Europe is falling behind the US on AI use at work, and the main reason is not technology, according to a new report.The report from Sifted, sponsored by London Business School, says Europe has an organisational gap rather than a tech gap.
Meta Ray-Ban Audio glasses are sleek and offer long battery life, but where will their AI go next?
Self-replicating AI worms could become a threat in 2027, but there are ways to spot them.
Z.ai’s GLM-Legal is a legal AI model built on GLM-5.3-Flash. It scores 19.2% all-pass on Harvey’s Legal Agent Benchmark, nearly 3x its base, at $0.15 per test. Access is by application.
Alibaba’s Qwen team has introduced QwenGyre, an elastic reinforcement learning RL framework for extreme-long-horizon agent training. QwenGyre is an online reinforcement learning system from Alibaba that trains coding agents through unmodified, black-box harnesses.
We checked 12 AI live translation tools, including Palabra.ai, Google Live Translate, DeepL Voice, Teams Interpreter and Wordly, against official pages for languages, latency, platforms and pricing.
TII has released Falcon-ASR, a 1.6B speech recognition model for Arabic with an Emirati focus. It reports 20.92% average WER on 6 Arabic test sets, but weights and an API are not yet available.
72 active sources
27 Turkish · 45 English
All sources
