Saazora
SaaS News
Oct 13, 2026 14 mins read 277 views

AI & SaaS News, Oct 13: Meta Muse Wonder, Gemini Skills

Meta's Muse Wonder costs $4 per million input tokens, Gemini starts replacing Gems with skills, Atlassian Cloud prices rise, IonQ buys Elisity for up to $1.3B, plus Synthesia, CodeRabbit, Ambi and Catalyst.

Quan Le
Quan Le
AI Tools Reviewer
AI & SaaS News, Oct 13: Meta Muse Wonder, Gemini Skills

Last checked October 13, 2026 · News roundup

Today in brief: Meta finally has a frontier model to sell. Muse Wonder went live on October 12 at $4 per million input tokens, while the stronger Muse Wonder Full sits behind a $50-a-month Muse Code plan. Google's Gemini app starts swapping Gems for skills today, Atlassian Cloud list prices rise today, and IonQ is paying up to $1.3 billion in stock for zero-trust security firm Elisity.

Below are eight stories that change what US users and software buyers pay, can use or should watch this week. Every price was checked against the vendor's own page or attributed to the outlet that reported it. If you are tracking renewals, our month-by-month SaaS price increase tracker has the full list for October.

October 13 at a glance
What changed, when, and who should care
Story
What changed
Who it affects
Meta Muse Wonder
Frontier model on Meta's API, $4 / $20 per 1M tokens (Oct 12)
Developers
Gemini skills
Skills reach the Gemini app from Oct 13; Gems conversion from Nov 17
Gemini users
Brad Smith on AI rules
Microsoft president calls for breach alerts and outside model checks (Oct 12)
AI buyers
Ambi
$26M for an always-listening AI assistant on iPhone, Mac and Watch (Oct 12)
Consumers
Atlassian Cloud
New list prices for Jira, Confluence, JSM and more (Oct 13 PT)
Teams on Atlassian
IonQ + Elisity
All-stock deal worth up to $1.3B (Oct 12)
Security buyers
Synthesia, CodeRabbit
Promo credits for a new video agent; agent price moved off the plan card (Oct 10)
Video and dev teams
Catalyst
$30M Series A led by Sequoia for AI trading agents (Oct 10)
Retail investors
Compiled by Saazora from Meta, Google, Atlassian, IonQ, TechCrunch, PricingSaaS and Crowdfund Insider, checked October 13, 2026.

Part 1 · AI

Meta Muse Wonder price: $4 per million tokens, Wonder Full on $50 plans

Meta released Muse Wonder and Muse Wonder Full on October 12, the first frontier models from Meta Superintelligence Labs. Meta pitches them at agent work: law, finance, health questions and coding, with two new features it calls Active Memory (carrying details across conversations) and Active Focus (retrying and checking in before important actions).

Access is split. Muse Wonder is on the Meta Model API now at $4 per million input tokens and $20 per million output tokens, with cached input at $0.20 and web search at $2.50 per 1,000 queries, according to VentureBeat. Muse Wonder Full is not sold through the API at all. You get it only with a Muse Code subscription on the $50-a-month Power Usage or $200-a-month Maximum Usage tier, Engadget reports. The cheaper $5 and $15 tiers don't include it.

On independent scoring, Artificial Analysis scores Wonder Full at 56, tied with Claude Sonnet 5.5 and two points behind Claude Opus 5.5, so "frontier" is fair, not "best." One privacy detail matters before you sign up: Muse Code subscriptions default to a Contributor mode that lets Meta train on your prompts and code. Turn it off in settings on day one. If you would rather build apps without writing code at all, our guide to AI app builders compares the no-code options.

Where Meta's new models land
Artificial Analysis Intelligence Index, higher is better (Meta models in orange)
Claude Opus 5.5 · 58
Muse Wonder Full · 56
Claude Sonnet 5.5 · 56
GPT-6 Astra · 53
Muse Wonder · 52
Grok 4.6 · 44
$4 / $20
Muse Wonder API, per 1M input / output tokens
$50+/mo
Muse Code tier needed for Wonder Full
Scores from Artificial Analysis as reported by VentureBeat on October 12, 2026; prices from VentureBeat and Engadget, October 12, 2026.

Gemini Gems to skills: what changes in the Gemini app from today

If you built custom Gems in Gemini, this is the week the replacement shows up. Google's Workspace admin help says skills begin rolling out in the Gemini app for all users on October 13, finishing by mid-November. During that window you can use both Gems and skills side by side, so nothing breaks today.

Skills are reusable instructions you call by typing "/" and a name. Unlike Gems, several can be stacked in one chat, and Gemini can run one automatically when a prompt matches. The catch for personal accounts: Google's skills help says they require users over 18, and TechRepublic reports they need Keep Activity switched on. Google also says most default Gem tools, including Canvas and Deep Research, don't work inside skills yet.

The deadline that matters is November 17. That's when an in-app notice, reported by 9to5Google and TechRepublic, says Google begins converting personal-account Gems into skills automatically. Business and enterprise Workspace accounts keep Gems until at least March 1, 2027. What to do now: open each important Gem, download its knowledge files, and rebuild it as a skill under Settings, Skills, "Create manually." Migrated skills arrive as inactive drafts, so plan to review them.

Gems to skills: the dates
Personal accounts move first; work and school accounts get months longer
Oct 5, 2026
Skills start rolling out to Google Workspace (Rapid Release done by Oct 12)
Oct 13, 2026
Today: skills start rolling out in the Gemini app for all users; Gems still work
Oct 19, 2026
Scheduled Release Workspace domains begin getting skills
Nov 17, 2026
Automatic conversion of personal-account Gems into skills begins
Mar 1, 2027
Earliest end of Gems for business, enterprise and nonprofit Workspace
Jun 1, 2027
Earliest end of Gems for education Workspace
Sources: Google Workspace Admin Help, Google Gemini blog (Sept 30) and TechRepublic (Sept 30), checked October 13, 2026.

Brad Smith says AI is less regulated than milk: the three rules Microsoft wants

Microsoft President Brad Smith used interviews with UK media, reported by TechCrunch and Engadget on October 12, to argue that the "moment has arrived" for binding AI safety rules. His line that grabbed headlines: a carton of milk and a washing machine face more safety rules than AI does.

Smith named three concrete ideas, per TechCrunch: a legal duty to alert an organization if rogue AI agents breached its systems, approval from outside evaluators before an AI product ships, and kill switches that let humans pause or slow a system. He called the White House's recent voluntary safety document a "good step" but not enough, and said the UK, not the US or the EU, is best placed to lead.

For US businesses, nothing changes in Copilot or Azure today. Read it as a preview of the questions your security team will be asking AI vendors next year: who tests the model before release, and will you be told if an agent goes rogue inside your account?

What is Ambi? The AI app that listens in the background raises $26 million

Ambi, founded by former Plaud executive Peter Mo, announced $26 million across several rounds on October 12, led by Luminous Ventures, TechCrunch reports. Its apps for iPhone, Mac and Apple Watch work like a meeting note-taker, but you can opt into all-day listening so the assistant picks up context and surfaces suggestions from your calendar, notes and the web.

Ambient capture is off by default, and Ambi says it doesn't store recordings and encrypts data, without naming the method. No price has been published. The open question is consent: Mo told TechCrunch it's on users to tell the people around them they're being recorded. Before you switch on all-day mode at work, check your company policy and your state's recording-consent rules, since some states require everyone in a conversation to agree.

Part 2 · SaaS & business

Atlassian price increase is live: what Jira and Confluence customers can still do

New Atlassian Cloud list prices took effect on October 13, 2026, Pacific time, for Jira, Confluence, Jira Service Management, Teamwork Collection, Atlassian Guard, Bitbucket Premium and Jira Product Discovery. Atlassian's FAQ doesn't publish percentages; partner Seibert puts the range at 3% to 10% depending on product, plan and seat count, with Guard at 3%.

What still protects you: any annual quote created before the deadline is honored at the old price until it expires, and a quote requested by a monthly customer can't run past November 11, 2026. What doesn't: plan upgrades and new products added from today use the new prices. Nonprofit, academic and open source discounts are unchanged, and Marketplace app prices aren't affected. We first flagged this deadline in our October 9 roundup; if you have an open quote, pay it before it lapses.

IonQ buys Elisity for up to $1.3 billion: what it means for zero-trust customers

Quantum computing company IonQ agreed on October 12 to acquire Elisity, a zero-trust network security vendor founded in 2019, in an all-stock deal worth up to $1.3 billion. That figure includes up to $150 million in earnout tied to Elisity's performance, and IonQ's share price for the deal is bounded between $33.07 and $49.61. Closing is expected in the fourth quarter, pending regulatory approval.

Elisity's IdentityGraph enforces least-privilege access for users, devices, workloads and, notably, AI agents, without replacing the underlying network. IonQ wants to bundle that with post-quantum encryption; CEO Niccolo de Masi said "Q-Day is coming much faster than predicted." Elisity says it has more than 100 customer deployments. If you're one of them, nothing changes before close, but ask your account team now about pricing and support terms for 2027 renewals.

Synthesia Syren and CodeRabbit Engineer: two quiet pricing moves to check

Synthesia added Syren, an AI video agent, to every plan on October 10, PricingSaaS reports. Synthesia's pricing page says each user gets 2,000 Syren-only promo credits for one month. Syren videos run up to three minutes and typically use 1,000 to 2,000 credits per finished minute, so the promo covers roughly one to two minutes of output. After that, Syren draws on the same credits as everything else: 500 a month on the free plan and 1,250 on the $29 Starter plan.

CodeRabbit, the AI code review tool, renamed its on-demand Agent add-on to CodeRabbit Engineer the same day and moved its $0.40-per-minute price off the plan card into the FAQ, according to PricingSaaS. Seats still list at $24 (Essentials), $48 (Team) and $72 (Advanced) per user per month. If you run these agents on autopilot, set a spending cap before the bill surprises you.

This week's price fine print
What each change costs a typical small team
Product
New pricing detail
Watch out for
Synthesia Syren
2,000 promo credits for 1 month; 1,000–2,000 credits per video minute
Then draws on plan credits (Starter: 1,250/mo)
CodeRabbit Engineer
$0.40 per agent minute, now listed only in the FAQ
On top of $24–$72 per seat
Atlassian Cloud
List prices up about 3%–10% from Oct 13 PT
Upgrades now billed at new rates
Sources: Synthesia and CodeRabbit pricing pages, PricingSaaS (Oct 10), Atlassian pricing FAQ and Seibert Group, checked October 13, 2026.

Catalyst raises $30 million for AI trading agents: what retail investors should know

San Francisco startup Catalyst raised a $30 million Series A led by Sequoia Capital, Crowdfund Insider reported on October 10, with Jump Trading, Lux Capital, Peak XV Partners, Coinbase Ventures and Premji Invest also in. Founder Justin Zheng is building a platform where autonomous agents design and run trading strategies for individual investors.

Users connect existing brokerage or crypto exchange accounts with API keys, then the agents backtest, size positions, hedge and place stop-losses inside risk limits the user sets. They work from financial news, live order-book data, social sentiment and macro history. No valuation or pricing was disclosed. Our take: the money says AI agents handling your money is coming fast. If you try one, give it a read-and-trade key with withdrawals disabled, start with a small balance, and set hard loss limits.

Searchers also ask

Is Meta Muse Wonder free?

No. Muse Wonder costs $4 per million input tokens and $20 per million output tokens on Meta's API. Muse Code subscriptions start at $5 a month, but the stronger Muse Wonder Full is only included on the $50 and $200 monthly tiers. Meta says Wonder will power its Muse agent, paid subscribers first.

Will I lose my Gemini Gems?

Not if you do nothing. Google says Gems convert to skills automatically, with personal-account conversions starting November 17, 2026, and work accounts later. Converted skills arrive as inactive drafts, and some Gem tools such as Canvas don't work in skills yet, so download knowledge files from any Gem you rely on now.

How much did Atlassian prices go up?

Atlassian doesn't publish a single percentage. Partner estimates put the October 13, 2026 increase at 3% to 10% depending on product, plan and seat count. Annual quotes created before the change are honored until they expire, and Atlassian's pricing tables show old and new list prices side by side.

Sources (accessed October 13, 2026): Meta Newsroom, Oct 12 · TechCrunch, Muse Wonder (Oct 12) · VentureBeat, Oct 12 · Engadget, Oct 12 · Meta, Muse Code subscriptions · Google Workspace Admin Help, Gems to skills · Gemini Apps Help · Google blog, Sept 30 · TechRepublic, Sept 30 · TechCrunch, Brad Smith (Oct 12) · Engadget, Oct 12 · TechCrunch, Ambi (Oct 12) · Atlassian Cloud pricing FAQ · Seibert Group, Sept 10 · IonQ, Oct 12 · PricingSaaS, Synthesia (Oct 10) · Synthesia pricing · PricingSaaS, CodeRabbit (Oct 10) · CodeRabbit pricing · Crowdfund Insider, Oct 10

Tags: #AI and SaaS news today
Quan Le
ABOUT THE AUTHOR

Quan Le

Quan Le is the founder and editor of Saazora, an independent site for news and reviews of SaaS and AI tools. He has worked in affiliate publishing since 2015, writing product reviews, verifying deals and running SEO for review sites. On Saazora he focuses on software for small businesses, creators and marketing teams, and evaluates tools by how well they fit real workflows: onboarding, integrations, pricing and long-term value, not the length of their feature list.

Read Quan's full profile · LinkedIn · contact@saazora.com

KEEP READING

Related Reviews