AgentStack Daily
by Nova & Alloy
Latest episode
Episode 125: Kimi K3.1 Pro Brings 1M-Token Context to OpenRouter
Episode 125: Kimi K3.1 Pro Brings 1M-Token Context to OpenRouter
About this podcast
Daily updates on agentic AI, local-first infrastructure, developer tools, model releases, and the systems behind modern AI workflows — everything builders and operators need to track the future of AI systems.
Episodes
The newest 20 of 127 episodes
- Episode 125: Kimi K3.1 Pro Brings 1M-Token Context to OpenRouter · 34 min · Episode 125AgentStack Daily leads with Kimi K3.1 Pro, now listed by OpenRouter with a 1M-token context window. Anthropic’s expanded agent sandbox, OpenAI’s Decisions... Show notes…
- Episode 124: MiniMax M3.1 Flash Targets Faster, Cheaper AI Agents · 33 min · Episode 124Today’s stories include MiniMax M3.1 Flash, llm-d 1.0 Cache-Aware Routing, SGLang’s new model-family support, and Mercury 3’s production-speed diffusion LLMs. Show notes…
- Episode 123: Reflection's Beam 1.0 Lands on OpenRouter as a 501B Coding MoE · 33 min · Episode 123Today on AgentStack Daily: Reflection's Beam 1.0 debuts on OpenRouter as a 501B coding MoE. Apple MLX runs a near-trillion-parameter Muse Spark on a Mac... Show notes…
- Episode 122: OpenAI Ships GPT-6, Gemini Goes Agentic, and EmbeddingGemma 2 Sees Everything · 33 min · Episode 122OpenAI ships GPT-6 globally with an Intelligent UI. Google turns Gemini into an objective-taking workplace agent, and DeepMind's EmbeddingGemma 2 puts four... Show notes…
- Episode 121: Mistral Large 4, EmbeddingGemma 2, and OpenTPU · 33 min · Episode 121Mistral Large 4 leads today’s news alongside EmbeddingGemma 2 and OpenTPU. Nova and Alloy examine documented capabilities and access, then cover agent... Show notes…
- Episode 120: GPT-6 Astra Ultrafast Runs 8x Faster on NVIDIA Blackwell · 32 min · Episode 120OpenAI ships GPT-6 Astra Ultrafast with 8x faster token generation on NVIDIA Blackwell. Plus: Aleph Alpha's 78B/3B-active Kolibri lands as open weights... Show notes…
- Episode 119: AgentStack Daily: RSA Puts AI Agents on the Identity Roster with Agent ID · 31 min · Episode 119RSA Puts AI Agents on the Identity Roster with Agent ID, Transformers 5.18 adds open-weight streaming speaker diarization, Firecrawl's MCP Server Turns Any... Show notes…
- Episode 118: Claude Sonnet 5.5 Arrives on OpenRouter With 1M Token Context Window · 35 min · Episode 118Claude Sonnet 5.5 lands on OpenRouter with a 1M token context window. Holo4 ships as open-weight computer-use agents across desktop, web, and Android. Show notes…
- Episode 117: GPT-6 Astra Cracks a 2005 Enigma as Mistral Releases a 675B Apache 2.0 Flagship · 34 min · Episode 117Mistral releases a 675B open-weight model on Apache 2.0 and ships Devstral 2 2512 for long-running coding agents. Show notes: https://tobyonfitnesstech.com/podcasts/episode-117/
- Episode 116: Claude Opus 5.5 Lands at $4 In, $20 Out, Meets GPT-6 Sol and Astra · 34 min · Episode 116Claude Opus 5.5 lands at $4 input and $20 output with Fable-class results, facing GPT-6 Sol and Astra in this episode's head-to-head. Show notes…
- Episode 115: AgentStack Daily: OpenAI Pitches Analytics to Tie ChatGPT and Codex to Business Outcomes · 33 min · Episode 115OpenAI Pitches Analytics to Tie ChatGPT and Codex to Business Outcomes, GitHub Copilot's September 14 batch adds model options, Sentry hook, and admin... Show notes…
- Episode 114: Prism-ML Ships Ternary Bonsai 2 27B: A 2-Bit Model Built For Local Hardware · 34 min · Episode 114Prism-ML ships Ternary Bonsai 2 27B, a 2-bit model engineered for local hardware. NVIDIA's Vera Rubin NVL72 tops MLPerf Inference v6.1 in its debut. Show notes…
- Episode 113: DeepSeek V4.1-Flash Packs Million-Token Memory Into FP4 Cache · 34 min · Episode 113DeepSeek V4.1-Flash lands with million-token context in an FP4 cache. Cohere releases a 218B open-weights translation model spanning 50 languages. Show notes…
- Episode 112: Google Antigravity, ChatGPT Mac with WebMCP, and MiniMax-Text-01's 4M Context Engine · 35 min · Episode 112Google Antigravity expands its agentic platform across desktop, CLI, and SDK. OpenAI brings Computer History and WebMCP site tools to ChatGPT on Mac. Show notes…
- Episode 111: AgentStack Daily: Agent Stack Release Readout: OpenClaw v2026.9.1 · 33 min · Episode 111Agent Stack Release Readout: OpenClaw v2026.9.1, Ling 3.0 Flash Fin lands on OpenRouter, a finance-focused MoE with 262K context, A Cheap Desktop 400GbE... Show notes…
- Episode 110: AgentStack Daily: Agent Stack Release Readout: OpenClaw v2026.8.2 · 33 min · Episode 110Agent Stack Release Readout: OpenClaw v2026.8.2, Qwen Team Open-Sources zg, a Local-First Search Layer for Agents, OpenClaw 2.0 dresses up an agent harness... Show notes…
- Episode 109: NVIDIA Jetson Orin Nano 2 Doubles Speed on New Silicon · 32 min · Episode 109Today's AgentStack Daily covers NVIDIA's Jetson Orin Nano 2 doubling throughput on new silicon, IBM's Granite 4.2 8B landing on OpenRouter with 131K... Show notes…
- Episode 108: AgentStack Daily: Cohere's Parse 5 Turns Scanned PDFs Into Clean Markdown · 33 min · Episode 108Cohere's Parse 5 Turns Scanned PDFs Into Clean Markdown, Claude, Codex, and Hermes Left 227 Unowned Install Commands in Corporate Docs, OpenAI and Thailand... Show notes…
- Episode 107: Codex Desktop's WebMCP Upgrade and Grok Bot's 24/7 Cloud Computer · 33 min · Episode 107Codex desktop rolls out WebMCP, Messages, Linux, and multi-repo review, while xAI's Grok Bot gives agents a 24/7 persistent cloud computer. Show notes…
- Episode 106: AgentStack Daily: Agent Stack Release Readout: OpenAI Codex rust-v0.149.0 · 35 min · Episode 106Agent Stack Release Readout: OpenAI Codex rust-v0.149.0, A new stealth reasoning model just landed on OpenRouter, Tencent's Hy-MT2-1.8B lands on OpenRouter... Show notes…
