What changed in AI today

• Series D valued Modal at $15.75B, led by Accel, per the company's announcement.
• Since its Series C, Modal's usage has tripled and revenue has grown more than 15x year over year.
• Its Sandboxes product reportedly reached a $300M run rate on its own, with close to 3 billion sandboxes spun up.
• Sandboxes serve background agents, reinforcement learning and vibecoding platforms, and reportedly scale to a million concurrent.

Why it matters for marketers: Content teams selling into agent builders can point to Modal's reported sandbox traction as evidence of demand for agent execution infrastructure.

• OpenAI's Joseph Kim predicts agents will handle investment idea generation and trading by 2027, per his keynote.
• Two Sigma's John Feminella says auto-mode classifiers handle routine calls and escalate only the most important requests to people.
• Anthropic's Keon Shahab says agents should propose order amendments for human approval, not make them directly.
• Bloomberg's Andrey Rybka urges deterministic controls outside the model, including kill switches, scoped credentials and audit trails.

Why it matters for marketers: Content and product teams building finance agents should plan approval steps and audit trails into the interface, not rely on model restrictions alone.

• Minos has moved from an internal playground to a full preprint, blog post and open dataset, released alongside the episode.
• Leash says its DNA-encoded library data could support tens of thousands of lab-in-the-loop simulations for agent training.
• Hermes, trained only on Leash's own data, reportedly keeps pace with Boltz on public benchmarks.
• An earlier Kaggle competition drew about 2,000 teams and roughly 4 billion measurements, and no team beat chance on an unseen library.

Why it matters for marketers: For AI and bio content teams, it offers a new open dataset for testing how agents reason about molecular structure and activity.

• Dyna-2.1 pairs new wheeled semi-humanoid hardware with an agentic system built around the Dyna-2 world-action model.
• Dyna reports a 99.4% success rate folding restaurant napkins over 24 straight hours, per co-founder Jason Ma's AIE talk.
• Robots reportedly fold thousands of napkins daily at Din Tai Fung, and can learn new tasks from as little as 45 minutes of demonstrations.
• Dyna-2 was pre-trained on over one million hours of human video and predicts future video to guide robot actions.

Why it matters for marketers: Content teams covering robotics get concrete benchmarks for hour-long autonomy, though the figures are company-reported and not independently verified.

• Available now via the API in AI Studio and Gemini platform, plus Google AI Ultra and Pro subscriber access in the Gemini app.
• Introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input 95% off.
• Google says Argon ranks #1 on the Vals Index, which measures performance across finance, coding, legal and tax work.
• Google reports Argon helped its own teams raise agentic code submissions by more than 35% in Antigravity.

Why it matters for marketers: Content and marketing teams using Gemini tools get a model Google positions for multi-step professional work, at a stated introductory price, though the Vals ranking is Google's own report.

• Announced Friday as "flows", a new multiagent orchestration feature built on Claude Managed Agents, presented at AI Engineer New York.
• Previously developers configured a coordinator and subagent roster; with flows, Claude writes an orchestration program that launches agents and manages information flow.
• Flows are available in Claude Console, and developers can still pass in pre-configured specialist sub-agents if they wish.
• Presenters said flows cost "much more" than a single agent but were "much more effective" at finding details; developers can set session budgets, not the model.

Why it matters for marketers: Teams running many agents should budget for higher token costs and use Claude Console graphing to audit what each sub-agent receives.

• Cache reads now cost $0.10 per million tokens on the Sonnet 5.5 API, according to ClaudeDevs.
• Standard input stays at $2 and output at $10 per million tokens.
• Anthropic estimates most agentic work becomes about 20% cheaper. Claude Code limits are unchanged.

Why it matters for marketers: Teams running agent workflows that reuse large cached context should re-run their API cost models, since cache reads drive the savings.

• AVEVA's Arti Garg says industrial AI now uses foundation, physical and agentic models, raising safety and predictability concerns.
• Garg describes a triple mandate for responsible AI: secure, efficient including environmentally, and preserving human safety and oversight.
• Garg says AI should augment rather than replace people in critical decision loops, with guardrails defining where systems can act.
• Garg is involved in an IEEE working group developing a standard for measuring AI's environmental impact, including water and carbon.

Why it matters for marketers: Content teams promoting industrial AI should expect buyers to ask about human oversight, governance and environmental impact, not only capability claims.

• Step 5 Preview costs $1 per million input tokens and $2.70 per million output tokens, per the sponsored item.
• Step 5 Preview reportedly scores near Claude Opus 5 (Medium) on Artificial Analysis's Intelligence Index, with a 1M-token context.
• Step 5 Preview is on OpenRouter and works in OpenCode, Kilo Code, Cline, omp and Hermes Agent; more integrations are planned.
• StepAudio 3 ASR ranks first in Artificial Analysis non-streaming transcription at 1.7% WER; Realtime ranks first in two conversational categories.

Why it matters for marketers: Teams building coding agents or budgeting for voice features can compare StepFun's per-token prices and benchmark rankings against current options.

• Stacklok, co-founded by Kubernetes creators Craig McLuckie and Joe Beda, pivoted into Kubernetes-based agentic solutions after starting in software supply chain security.
• Mecatl, an open source cloud-native harness, began on GitHub in June and keeps the agent loop separate from client, model provider, state and execution.
• Session management and memory move out of local JSONL files into manageable systems that enterprises can govern centrally.
• Stacklok's AI Gateway, not yet open sourced, handles access control, budgets, reporting and provider routing rather than task-based model selection.

Why it matters for marketers: Content and marketing teams at regulated enterprises may need to weigh centrally governed agent infrastructure against desktop-first coding tools when planning AI workflows.