NewsPodLM Daily Podcast 14 Feb 2026

Anthropic Shatters Funding Records with Thirty Billion Dollar Raise

Anthropic has closed the largest private funding round in AI history, raising thirty billion dollars in Series G financing at a staggering three hundred and eighty billion dollar post-money valuation. The round was led by GIC and Coatue, with participation from D. E. Shaw Ventures, Dragoneer, Founders Fund, Sequoia Capital, BlackRock, and Goldman Sachs among others. The company reports annualised revenue of approximately fourteen billion dollars, with enterprise customers spending over one hundred thousand dollars annually growing sevenfold in the past year.

OpenAI Launches Lightning-Fast Coding Model on Cerebras Chips

OpenAI released GPT-5.3-Codex-Spark, a streamlined version of its Codex model optimised for real-time coding. The model generates over one thousand tokens per second by running on Cerebras' Wafer Scale Engine 3, marking OpenAI's first production deployment away from NVIDIA hardware. Available as a research preview to ChatGPT Pro users, Spark maintains strong performance on software engineering benchmarks while completing tasks dramatically faster than its full-sized sibling.

DeepMind's Aletheia Autonomously Solves Open Maths Problems

Google DeepMind has unveiled Aletheia, an AI agent powered by Gemini Deep Think that moves beyond competition-level mathematics to conduct autonomous research. The system uses a three-part architecture of generator, verifier, and reviser to produce and validate mathematical proofs. In testing against seven hundred open problems from the Erdos Conjecture database, Aletheia independently solved four, with one leading to a published paper.

MiniMax Delivers Near-Frontier Performance at a Fraction of the Cost

Chinese AI lab MiniMax released M2.5 and M2.5-Lightning, open-source models that score eighty point two percent on SWE-Bench Verified, within striking distance of Claude Opus 4.6. At just two dollars and forty cents per million output tokens, M2.5 costs one-tenth to one-twentieth of competing frontier models.

AI Brain Scanner Achieves Ninety-Seven Percent Accuracy

Researchers at the University of Michigan have published results in Nature Biomedical Engineering for Prima, an AI system that diagnoses brain MRI scans in seconds with up to ninety-seven point five percent accuracy across more than fifty neurological conditions, trained on over two hundred thousand MRI studies.