Next Tool

Archives
Log in
Subscribe
July 12, 2026

GPT-5.6 lands, Meta joins the coding race, and Google starts labeling AI ads

Subject: GPT-5.6 lands, Meta joins the coding race, and Google starts labeling AI ads A/B alternates: - 5 AI things worth your 4 minutes this week - New tools from OpenAI, Meta, and Google — the honest take


TL;DR: OpenAI shipped GPT-5.6 in three sizes (Luna/Terra/Sol) and a new agentic API surface. Meta launched Muse Spark 1.1 for agentic coding at aggressive pricing. Google will start labeling ads made with AI. And Anthropic's Fable 5 still beats GPT-5.6 on at least one coding benchmark. Read on for the details that actually matter.

OpenAI's new flagship came out hot. So did Meta's answer. So did Google's ad policy. We sifted through the noise so you can decide what to actually try this week.


1. GPT-5.6 family (Luna, Terra, Sol) — OpenAI's newest flagship model family hit general availability on July 9, 2026 in three sizes priced per 1M input/output tokens: Luna $1/$6, Terra $2.50/$15, Sol $5/$30. Why it matters: Sol scored 53.6 on the "Agents' Last Exam" benchmark across 55 professional fields — 13.1 points above Anthropic's Claude Fable 5 at high reasoning, and 11.4 points above it at medium reasoning at roughly a quarter of the cost. Honest take: On SWE-Bench Pro, Fable 5 still wins (80% vs. 64.6% for Sol) — so for hard coding work today, Anthropic's model is still the one to beat. New API features (programmatic tool calling, native multi-agent subagents) are worth a look for builders. Source: https://simonwillison.net/2026/Jul/9/gpt-5-6/

2. Meta Muse Spark 1.1 — Meta publicly launched Muse Spark 1.1 on July 9, 2026 — a multimodal model built for agentic coding, large code migrations, and bug fixes that compete with OpenAI's Codex and Anthropic's Claude Code. Pricing (per Reuters via TechCrunch): $1.25 / $4.25 per 1M input/output tokens. Why it matters: That's roughly in line with Claude Haiku 4.5 and GPT-5.6 Luna — Meta is undercutting premium-tier coding models by being cheaper than Anthropic's flagship on the same workloads. Honest take: Meta is late here. If you've already got a Claude Code or Cursor workflow you trust, no reason to switch. If you don't, the pricing makes Spark 1.1 worth a 30-minute benchmark against your real tasks. Source: https://techcrunch.com/2026/07/09/meta-enters-the-crowded-ai-coding-battle-with-muse-spark-1-1/

3. Google starts labeling AI-made ads — As of July 9, 2026, Google is rolling out a disclosure feature that flags ads containing AI-generated or digitally altered content. Until now, this disclosure was only required for election ads. Why it matters: If you run paid acquisition or you buy ads to appear in front of AI-search audiences, your creative will get tagged. Plan copy and asset review accordingly. Honest take: Good move for users, mild friction for advertisers. Not a strategy change — a labeling change. Source: https://techcrunch.com/2026/07/09/google-will-now-disclose-which-ads-are-made-with-ai/

4. OpenAI hiring a "Families" PM — On July 11, 2026, TechCrunch reported OpenAI is hiring a dedicated product manager in San Francisco to build experiences for families, caregivers, and older adults across ChatGPT. Why it matters: Three years after launch, ChatGPT is shifting from "individual knowledge worker" to "household utility." Expect family plans, kid profiles, and caregiver controls inside the next 2–4 quarters. Honest take: Useful if you share ChatGPT with your kids or parents today. For builders: any new product surface = a new API or app surface eventually. Watch the changelog. Source: https://techcrunch.com/2026/07/11/openai-bets-on-families-as-chatgpt-goes-deeper-into-households/

5. Gradium raises $100M seed for AI voice — Paris-based AI voice startup Gradium closed a $100M seed round backed by Nvidia, announced July 9, 2026. The company plans to open a Bay Area office and compete for talent "at the heart of the world's leading AI ecosystem." Why it matters: Voice is the next interface layer — if you ship anything with TTS or voice agents, the realistic-tier options are about to get way better and cheaper. Honest take: Big seed, but a seed is a seed. Wait for a public API or v1 product before you rebuild your stack on it. Source: https://techcrunch.com/2026/07/09/paris-based-ai-voice-startup-gradium-raises-100m-seed-backed-by-nvidia/


That's it for this week. Forward this to one friend who'd find it useful — that single share is the #1 way this newsletter grows.

— Robert Next Tool — What's next in AI productivity.

Affiliate disclosure: Some links in this issue may earn Next Tool a commission at no extra cost to you. We only recommend tools we'd use ourselves.

Don't miss what's next. Subscribe to Next Tool:
← Newer GPT-5.6 just beat Claude Opus — and Claude Code ships 33k tokens before your prompt Older → GPT-5.6 lands in three sizes, Meta joins the coding wars, and an AI agent raises its own 00M
sauteri.com
Powered by Buttondown, the easiest way to start and grow your newsletter.