The record for this issue
Gemini 4 Argon switches on today.
43 claims in the draft, each tested by a second agent against the page it cites, on Oct. 27, 2026. 42 held; 1 did not and were fixed or cut before this went out.
-
01
dekconfirmed
In Gemini Enterprise, Gemini 4 Argon defaults to on from Oct. 27 in the web and mobile apps and in Workflow Builder
The page says “Oct. 13: GA "for the Gemini Enterprise web app, mobile app, and AI developer tools"; "Turned off by default. The toggle defaults to on beginning October 27, 2026." Oct. 14 (Workflow Builder): "Turned off by default.…”
Gemini Enterprise release notes, at the passage · whole page
-
02
dekconfirmed
Administrators can act in Feature Management
The page says “select your app ... Configurations ... Feature Management tab ... Model availability”
-
03
summaryconfirmed
AlphaEvolve's default model moved Monday (Oct. 26)
The page says “On October 26, 2026, Gemini 4 Argon will replace Gemini 3.8 Flash as the default model for new AlphaEvolve experiments”
Gemini Enterprise release notes, at the passage · whole page
-
04
The storyconfirmed
From Oct. 27, Argon is set to be on by default in the Gemini Enterprise web and mobile apps
The page says “App availability: Turned off by default. The toggle defaults to on beginning October 27, 2026”
Gemini Enterprise release notes, at the passage · whole page
-
05
The storyconfirmed
Google's admin documentation was updated Oct. 26
The page says “Last updated 2026-10-26 UTC”
-
06
The storyconfirmed
The docs list Argon's toggle as turned off in the global, US and EU locations
The page says “Turned off in global, us, and eu (defaults to on beginning October 27, 2026)”
-
07
The storyconfirmed
Quote: "defaults to on beginning October 27, 2026." (6 words, attributed to Google's admin documentation)
The page says “(defaults to on beginning October 27, 2026)”
-
08
The storyconfirmed
The same date applies to Argon in Workflow Builder and workflow agents
The page says “Oct. 14: "Gemini 4 Argon is also available in Workflow Builder workflow agents." "Turned off by default. The toggle defaults to on beginning October 27, 2026."”
Gemini Enterprise release notes, at the passage · whole page
-
09
The storyconfirmed
Argon became generally available in Gemini Enterprise on Oct. 13, switched off at first (labelled background)
The page says “October 13, 2026 ... Generally available (GA) ... Turned off by default”
Gemini Enterprise release notes, at the passage · whole page
-
10
The storyconfirmed
Google describes Argon as built for long-horizon work needing deep reasoning, such as code migrations and frontier research
The page says “designed for those long-horizon workflows that require deep reasoning and structural precision, including code migrations and frontier research”
-
11
The storyconfirmed
Path: the Gemini Enterprise page in the Cloud console, then the app, Configurations, the Feature Management tab, where Model availability sets which models users can pick
The page says “Gemini Enterprise page ... click the name of the app ... Configurations ... Feature Management tab ... Model availability”
-
12
The storyconfirmed
The model selector has its own toggle
The page says “ensure Enable model selector is turned on”
-
13
The storyconfirmed
For some GA models, such as Gemini 3.5 Flash, the toggle can't be turned off
The page says “For some models in General Availability (GA) in the Gemini Enterprise app, such as Gemini 3.5 Flash, you can't turn the toggle off”
-
14
The story, step 1confirmed
Oct. 13: Argon generally available, toggle off
The page says “October 13, 2026 ... Generally available (GA) ... Turned off by default”
Gemini Enterprise release notes, at the passage · whole page
-
15
The story, step 2confirmed
Oct. 22: AlphaEvolve default change announced
The page says “October 22, 2026 — Announcement: On October 26, 2026, Gemini 4 Argon will replace Gemini 3.8 Flash as the default model for new AlphaEvolve experiments”
Gemini Enterprise release notes, at the passage · whole page
-
16
The story, step 3confirmed
Oct. 26: AlphaEvolve defaults to Argon
The page says “On October 26, 2026, Gemini 4 Argon will replace Gemini 3.8 Flash as the default model”
Gemini Enterprise release notes, at the passage · whole page
-
17
The story, step 4confirmed
Oct. 27: App and Workflow Builder toggles default on
The page says “The toggle defaults to on beginning October 27, 2026. (Oct. 13 app entry and Oct. 14 Workflow Builder entry)”
Gemini Enterprise release notes, at the passage · whole page
-
18
The morning in four, item 1unsupported
OpenAI released GPT-6.1 Terra on Oct. 26, as Simon Willison reported
-
19
The morning in four, item 1confirmed
Terra priced at $1 per million input tokens and $5 per million output
The page says “GPT-6.1 Terra is priced at $1/million input tokens and $5/million output”
-
20
The morning in four, item 1confirmed
Almost half GPT-6.1 Sol's price
The page says “almost half the price of GPT-5.6 Terra and GPT-6.1 Sol”
-
21
The morning in four, item 2confirmed
Since Oct. 26, new AlphaEvolve experiments default to Gemini 4 Argon and bill at its rates
The page says “From that date, new experiments that use the default model are billed at Gemini 4 Argon token prices instead of Gemini 3.8 Flash token prices”
Gemini Enterprise release notes, at the passage · whole page
-
22
The morning in four, item 2confirmed
Name Gemini 3.8 Flash to keep it
The page says “specify gemini-3.8-flash in the models field”
Gemini Enterprise release notes, at the passage · whole page
-
23
The morning in four, item 2confirmed
Existing experiments keep their models
The page says “Experiments created before October 26, 2026 keep the model recorded in their configuration”
Gemini Enterprise release notes, at the passage · whole page
-
24
The morning in four, item 3confirmed
Nanbeige LLM Lab's Nanbeige4.3-33B is on Hugging Face under Apache 2.0
The page says “License: Apache 2.0; Nanbeige LLM Lab”
-
25
The morning in four, item 3confirmed
524,288-token context
The page says “context length 524,288 tokens (512K)”
-
26
The morning in four, item 3confirmed
Aimed at agentic coding and office tasks, per its model card
The page says “excels at complex coding and professional cowork tasks ... agentic coding and office tasks”
-
27
The morning in four, item 4confirmed
Google's AI Quests added Polar Peaks on Oct. 26, a neuroscience quest
The page says “October 26, 2026 ... Polar Peaks ... how to map the human brain using artificial intelligence”
-
28
The morning in four, item 4confirmed
For students aged 11 to 14
The page says “students ages 11 to 14”
-
29
The morning in four, item 4confirmed
More than 400,000 quests completed (quests, not students)
The page says “More than 400,000 quests have been completed to date, localized to 10 languages”
-
30
The storyconfirmed
Since Oct. 22, Gemini Enterprise admins can set single-user quota limits on the Usage & Spending > Usage page
The page says “October 22, 2026: Gemini Enterprise administrators can configure single-user quota limits on the Usage & Spending > Usage page”
Gemini Enterprise release notes, at the passage · whole page
-
31
The storyconfirmed
So that one user can't consume an edition's whole shared quota pool
The page says “prevent an individual user from using all of an edition's shared subscription quota pool”
Gemini Enterprise release notes, at the passage · whole page
-
32
The storyconfirmed
Google lists the Standard, Plus and Standard Emerging Market editions
The page says “Standard, Plus, and Standard Emerging Market editions”
Gemini Enterprise release notes, at the passage · whole page
-
33
The storyconfirmed
Google's note doesn't suggest a level
The page says “The Oct. 22 entry's sentences: configure limits; "You can enable a single-user quota limit and set a quota multiplier…"; "This feature is generally available (GA)." No value is recommended”
Gemini Enterprise release notes, at the passage · whole page
-
34
The storyconfirmed
The limit is set as a multiplier
The page says “set a quota multiplier for each edition and location”
Gemini Enterprise release notes, at the passage · whole page
-
35
The storyconfirmed
Quote: OpenAI calls Terra "a significant step up from GPT‑5.6 Terra," as quoted by Willison (8 words)
The page says “a significant step up from GPT‑5.6 Terra”
-
36
The storyconfirmed
OpenAI's Ted Sanders wrote it in a Hacker News comment that Willison quoted
The page says “OpenAI's Ted Sanders provided some useful advice in a comment on Hacker News”
-
37
The storyconfirmed
Sanders put Terra probably ahead of Astra and Sol on technical communication
The page says “I'd say Terra is probably better than Astra and Sol at technical communication”
-
38
The storyconfirmed
Similar on vision and computer use, but much faster
The page says “similar on vision and computer use (but much faster!)”
-
39
The storyconfirmed
Quote: "a bit worse on most other things." (7 words)
The page says “and a bit worse on most other things”
-
40
The storyconfirmed
Quote: for subscribers, "it's not a bombshell release." (5 words)
The page says “For subscribers ... it's not a bombshell release”
-
41
The storyconfirmed
Interpretation: by the vendor's own account, Terra's case is price and speed on visual, computer-use and writing tasks, not a general upgrade
The page says “a nice midrange option if you want something cheaper than Sol but a lot smarter than Luna (especially for vision / computer use / communication)”
-
42
What to watch, item 1confirmed
Today Argon toggles default on in the web and mobile apps and Workflow Builder, for the global, US and EU locations
The page says “Turned off in global, us, and eu (defaults to on beginning October 27, 2026); release notes Oct. 13/14: The toggle defaults to on beginning October 27, 2026”
-
43
What to watch, item 2confirmed
From Oct. 26, new experiments that use the default model bill at Gemini 4 Argon rates
The page says “From that date, new experiments that use the default model are billed at Gemini 4 Argon token prices”
Gemini Enterprise release notes, at the passage · whole page