Models & Agents
by Patrick · Nerra Network
Latest episode
Ep 234: Vercel’s UI SDK supplies agents with 50 behaviors for…
Ep 234: Vercel’s UI SDK supplies agents with 50 behaviors for…
About this podcast
Your daily briefing on AI models and agents: new releases from the frontier labs, open-weight drops, agent frameworks, benchmarks, pricing, and practical tools you can use the same day — with long-running program tracking so you always know where the big stories stand. For developers, builders, and AI practitioners.
AI Disclosure: This podcast is curated by Patrick but uses AI-generated voice synthesis for audio production.
Episodes
The newest 20 of 234 episodes
- Ep 234: Vercel’s UI SDK supplies agents with 50 behaviors for building accessible components… · 10 min · Season 1 Episode 234Models & Agents Vercel’s UI SDK supplies agents with 50 behaviors for building accessible components across five frameworks. What You Need to Know: Vercel Labs released an…
- Ep 233: Gemini 4 Hydrogen ships at $1.50/$7.50 per million tokens and powers the free tier plus… · 10 min · Season 1 Episode 233Models & Agents Gemini 4 Hydrogen ships at $1.50/$7.50 per million tokens and powers the free tier plus Google Search. What You Need to Know: Google released Gemini 4 Hydrogen…
- Ep 232: Google's Gemini Decision 1 scores up to 1,000 multimodal questions in one 150-250 ms pass… · 9 min · Season 1 Episode 232Models & Agents Google's Gemini Decision 1 scores up to 1,000 multimodal questions in one 150-250 ms pass at $0.10 per million input tokens with zero output cost. What You Need…
- Ep 231: A new 0.8B on-device model doubles agentic benchmark scores while keeping all data private. · 12 min · Season 1 Episode 231Models & Agents A new 0.8B on-device model doubles agentic benchmark scores while keeping all data private. What You Need to Know: Tether AI released BotPsy, an 0.8B model…
- Ep 230: Anthropic cut Fable 5.2 prices 20 percent while raising its FrontierCode 2.0 score from… · 9 min · Season 1 Episode 230Models & Agents Anthropic cut Fable 5.2 prices 20 percent while raising its FrontierCode 2.0 score from 32.4 to 49.1 percent at xhigh effort. What You Need to Know: Anthropic…
- Ep 229: LG just opened free API access to its 750B K-EXAONE 2.5 model for agentic workloads. · 14 min · Season 1 Episode 229Models & Agents LG just opened free API access to its 750B K-EXAONE 2.5 model for agentic workloads. What You Need to Know: LG AI Research released free hosted access to…
- Ep 228: Ten open MoE models expose routing choices that swing sparsity from 4x to 128x and change… · 9 min · Season 1 Episode 228Models & Agents Ten open MoE models expose routing choices that swing sparsity from 4x to 128x and change which hardware setups actually fit. What You Need to Know: A verified…
- Ep 227: Ling-3.2-flash ships a 560B/25B multimodal MoE with 1M context and OpenAI-compatible APIs… · 9 min · Season 1 Episode 227Models & Agents Ling-3.2-flash ships a 560B/25B multimodal MoE with 1M context and OpenAI-compatible APIs for immediate use. What You Need to Know: Ant Ling released…
- Ep 226: Gemini Omni 1.2 Flash handles longer video clips on edge devices at lower cost than prior… · 11 min · Season 1 Episode 226# Models & Agents Gemini Omni 1.2 Flash handles longer video clips on edge devices at lower cost than prior Gemini versions. What You Need to Know: Google released Gemini Omni…
- Ep 225: Together Compiler lets agents re-optimize inference on the fly and delivers up to 2.6… · 18 min · Season 1 Episode 225Models & Agents Together Compiler lets agents re-optimize inference on the fly and delivers up to 2.6 times higher throughput. What You Need to Know: OpenAI released…
- Ep 224: OpenAI is now shipping prototype hardware and CAD tools built with GPT-6 Astra in under a… · 19 min · Season 1 Episode 224Models & Agents OpenAI is now shipping prototype hardware and CAD tools built with GPT-6 Astra in under a week. What You Need to Know: OpenAI released a set of engineering…
- Ep 223: An LLM edited just 0.016 percent of a small model's weights to raise its math benchmark… · 11 min · Season 1 Episode 223Models & Agents An LLM edited just 0.016 percent of a small model's weights to raise its math benchmark score from 7.9 to 48.1 percent. What You Need to Know: MIT CSAIL…
- Ep 222: Apriel's 27B open model scores 98.9 on Big Bench Audio while adding native speech and tool… · 17 min · Season 1 Episode 222Models & Agents Apriel's 27B open model scores 98.9 on Big Bench Audio while adding native speech and tool calling. What You Need to Know: Apriel Voice 27B Preview unifies…
- Ep 221: DeepSeek’s new super-agent routes GitHub and Feishu requests into parallel worker… · 11 min · Season 1 Episode 221Models & Agents DeepSeek’s new super-agent routes GitHub and Feishu requests into parallel worker sessions, cutting p95 task time from 29.5 to 15.7 minutes. What You Need to…
- Ep 220: AI labs gave publishers one standardized signal to block future training scrapers across… · 10 min · Season 1 Episode 220Models & Agents AI labs gave publishers one standardized signal to block future training scrapers across OpenAI, Google, Microsoft and Anthropic at once. What You Need to Know…
- Ep 219: Samaya Sterling beats Claude Fable 5.1 on FrontierFinance at one-eleventh the cost. · 9 min · Season 1 Episode 219Models & Agents Samaya Sterling beats Claude Fable 5.1 on FrontierFinance at one-eleventh the cost. What You Need to Know: Samaya AI released Sterling, a LoRA post-trained…
- Ep 218: Gemini 4 Carbon scores 88.2 percent on vulnerability finding and leads the Artificial… · 11 min · Season 1 Episode 218Models & Agents Gemini 4 Carbon scores 88.2 percent on vulnerability finding and leads the Artificial Analysis Intelligence Index at 59. What You Need to Know: Google DeepMind…
- Ep 217: Genie 4 now gives builders real-time physics, reasoning, and audio in interactive… · 14 min · Season 1 Episode 217Models & Agents Genie 4 now gives builders real-time physics, reasoning, and audio in interactive simulations they can test today. What You Need to Know: DeepMind has begun…
- Ep 216: OpenAI’s GPT-6.1 Terra delivers faster computer-use agents at half the prior token price. · 12 min · Season 1 Episode 216Models & Agents OpenAI’s GPT-6.1 Terra delivers faster computer-use agents at half the prior token price. What You Need to Know: OpenAI began rolling out GPT-6.1 Terra to…
- Ep 215: Kolibri activates 3.46 billion parameters per token while landing on the Pareto frontier… · 12 min · Season 1 Episode 215Models & Agents Kolibri activates 3.46 billion parameters per token while landing on the Pareto frontier for English and German quality versus serving cost. What You Need to…
