RuntimeWire

Real-time startup intelligence for the AI economy. Funding, model launches, infra shifts, and founder moves — the signals that matter.

Latest stories — page 12

• grok-4.3 edges gpt-5.4-nano on execution, not flash

  This was close on aggregate, but grok-4.3 wins because it made the fewer costly mistakes in structured-output work. gpt-5.4-nano was sharper on tone and regex edge cases, yet it gave back those gains by breaking instructions where precision mattered more.

• Tesla hacker Yoni Ramon brings Pi out of stealth with $35M for AI security

  Pi is valued at $100 million and counts Navan as an early customer, while Forbes reports xAI is also using the system.

• Mike Krieger turns Anthropic's Fable 5 launch into a product test

  Anthropic says Fable 5 routes under 5% of sessions to Opus 4.8, while Mythos 5 keeps higher-risk capability behind trusted access.

• Instawork turns its gig marketplace into a robot-training data line

  Instacore puts five cameras and a compute backpack on workers to capture commercial tasks for AI labs, with customers still unnamed.

• Gamow Labs Turns a Missed NICU Diagnosis Into a Genomics Startup

  The Boulder company is building an AI interpretation layer for whole genomes, after its founder says standard testing missed his son's fatal disorder.

• Nagish becomes Rylo and raises $85 million for AI call captioning

  Co-founder Tomer Aharoni is taking a school project into a broader accessibility platform backed by General Catalyst and Canaan.

• Luma Ray 3.2 steamrolls AnimateDiff

  Across both prompt-following tests, Luma Ray 3.2 Image to Video wasn’t just better than AnimateDiff—it was operating in a different league. AnimateDiff could gesture at mood; Luma delivered the actual scene, action, and camera logic the prompts asked for.

• Life Biosciences tests cellular reprogramming in a human eye

  David Sinclair's longevity company has moved ER-100 from animal work into a small FDA trial for glaucoma and optic nerve disease.

• Fibo Bbq Preview beats Bagel on image direction

  Bagel steals one poster task, but Fibo Bbq Preview wins the matchup where it matters: prompt control, scene construction, and mood. On aggregate, B is the more reliable image model and the clear overall pick.

• Modulus is pitched as a multiplayer IDE for engineers and AI agents, with key details still undisclosed

  A third-party launch post says Modulus puts human engineers and AI agents in the same real-time workspace. The available source does not disclose founders, pricing, availability, deployment, integrations or security controls.

• Anthropic launches Claude Fable 5 with a gated Mythos 5 for cyber use

  The new model is priced at $10 per million input tokens and $50 per million output tokens, with some requests routed to Opus 4.8.

• Illyas Khan's Quantinuum IPO turns a quantum software bet into a $2.2 billion stake

  Quantinuum raised $1.68 billion at a valuation above $15.6 billion while reporting just $30.9 million in 2025 revenue.

• Decagon puts its CX agents on a self-improvement loop

  Duet Autopilot turns live support conversations into proposed agent updates, but Decagon is keeping humans in the approval seat.

• Google ships Gemini 3.5 Live Translate across consumer, enterprise and developer tools

  The audio model streams speech-to-speech translation across 70-plus languages, with Meet access limited to private preview this month.

• ICEYE's reported EUR10 billion valuation turns radar satellites into Europe's space benchmark

  Bloomberg's valuation report leaves the deal mechanics unclear, but it puts a 2014 Finnish SAR operator in a new European space tier.

• grok-4.3 vs DeepSeek-V4-Flash: Precision Beats Polish

  grok-4.3 takes this matchup by being the stricter, cleaner finisher on structured-output tasks, while DeepSeek-V4-Flash wins the one audience-sensitive writing test. The scoreline is close, but the deciding errors are the kind that matter in production.

• Standard Bots raises $200M to manufacture robots in the US

  Bloomberg reported the financing, but the round's investors, valuation, close date and factory plan remain undisclosed in the accessible record.

• Alex Xiao's Evotrex raises $30M for a hybrid RV that can live off-grid

  The Anker-backed Los Angeles company plans 2027 sales of the PG5, a battery trailer with an onboard gas engine to recharge it.

• Report: Tim Cook sets September handoff to John Ternus after WWDC AI pitch

  Apple's own WWDC pages highlight Siri AI and child-safety updates, while the CEO succession claim rests on an X post and secondary sourcing.

• Zepto's IPO filing shows Aadit Palicha's ad bet is outpacing overall revenue growth

  The Indian quick-commerce company filed for an IPO with revenue doubling, losses widening, and public investors still waiting on valuation clarity.

• Harness-1 researchers say a 20B open search agent beat GPT-5.4 on recall

  The UIUC, UC Berkeley and Chroma project shifts search memory from the model context window into a structured software environment.

• Mercor's Brendan Foody calls out Sequoia's dual-price valuation play

  Foody says some AI rounds are marketed at headline prices that do not match where lead investors put most of their money.

• Marey Realism V1.5 Beats AnimateDiff Where It Counts

  AnimateDiff is the steadier clip-maker, but Marey Realism V1.5 is the better prompt reader and the more convincing filmmaker. Across both tests, it delivered the details, atmosphere, and camera language the prompts actually asked for.

• Intuned bets browser automation still needs code, just less of it

  The YC S22 company says Intuned Agent writes and repairs Playwright automations, while leaving revenue, customer names and funding terms undisclosed.