Release-claim watchlist · Method 1.0.0

Announced. But available?

AI Release Tracker follows product release claims in our published news archive. Use it to revisit an announcement, inspect its recorded sources, and decide what needs checking next.

Snapshot assembled . Latest matching article: . This is not a source-access or availability-check timestamp.

Release-claim articles
223
Availability unchecked
223
7-day checks due
210
30-day checks due
84
Availability not independently checked. Every entry is an automatically selected claim, including stories that say “available now.” Unknown availability does not mean unavailable. No access tests or completed follow-ups are recorded here.

Where every claim stands

Follow-up status

Suggested check dates fall 7 and 30 days after each article was published. “Due” means the date has passed; no check has been completed for any claim.

7-day suggested checks

223 claims
  • Due; not completed 210
  • Upcoming; not scheduled 13
  • Date unknown 0

30-day suggested checks

223 claims
  • Due; not completed 84
  • Upcoming; not scheduled 139
  • Date unknown 0

Watchlist mentions

Articles mentioning
  • OpenAI50
  • Google / DeepMind43
  • NVIDIA36
  • Anthropic32
  • Meta27
  • Alibaba / Qwen21
  • Microsoft16
  • DeepSeek13

Mentions, not confirmed ownership: one article can mention several organizations.

The watchlist

Find a release claim

“Major” means names on our explicit watchlist, not a verified importance ranking. Organization matches are mentions, not confirmed product ownership. Dates are UTC.

Reset filters

223 matching articles · page 7 of 12. Export includes every matching record, not just this page.

Article published · Product

Google rolls out Gemini AI phone calling feature for Pixel 11 users

Availability not independently checked

Google is launching an early experiment on Pixel 11 that allows Gemini to make local business calls on behalf of users, handling hold times and conversations with a live transcript option.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “rolls out”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • theverge.comhttps://www.theverge.com/ai-artificial-intelligence/1000116/google-gemini-business-phone-callsRecorded: Unknown · Article source field
  • theverge.comhttps://www.theverge.com/ai-artificial-intelligence/1000116/google-gemini-business-phone-callsRecorded: · Source history

Article published · Product

Black Forest Labs releases FLUX 3 Action open-weights robotics model

Availability not independently checked

Black Forest Labs has released FLUX 3 Action, a 7-billion-parameter open-weight World Action Model that claims to outperform larger competitors on NVIDIA's RoboLab-120 benchmark while offering faster inference speeds for practical robotics deployment.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • venturebeat.comhttps://venturebeat.com/infrastructure/black-forest-labs-debuts-flux-3-action-an-open-weights-ai-robotics-model-that-tops-the-leaderboard-at-half-the-size-of-its-competitionRecorded: Unknown · Article source field
  • venturebeat.comhttps://venturebeat.com/infrastructure/black-forest-labs-debuts-flux-3-action-an-open-weights-ai-robotics-model-that-tops-the-leaderboard-at-half-the-size-of-its-competitionRecorded: · Source history

Article published · Product

Google Antigravity SDK adds support for local AI models

Availability not independently checked

Google has updated the Antigravity SDK to support local AI workflows, enabling developers to run agentic capabilities offline using models like Gemma 4 26B A4B via LiteRT.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “adds”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • developers.googleblog.comhttps://developers.googleblog.com/introducing-support-for-local-ai-models-in-the-antigravity-sdk/Recorded: Unknown · Article source field
  • developers.googleblog.comhttps://developers.googleblog.com/introducing-support-for-local-ai-models-in-the-antigravity-sdk/Recorded: · Source history

Article published · Innovation

Google DeepMind adds persistent private memory to AI compute platform

Availability not independently checked

Google DeepMind has updated its Private AI Compute architecture to support persistent, cross-device AI memory using server-side secure enclaves and device-held encryption keys.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “adds”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • deepmind.googlehttps://deepmind.google/blog/advancing-private-ai-compute-with-secure-server-side-memory/Recorded: Unknown · Article source field
  • deepmind.googlehttps://deepmind.google/blog/advancing-private-ai-compute-with-secure-server-side-memory/Recorded: · Source history

Article published · Product

Nvidia releases Isaac ROS 5.0 with new AI agent tools for robotics developers

Availability not independently checked

Nvidia has released Isaac ROS 5.0, an update to its open-source robotics software that adds tools for AI agents, expanded ROS support and new capabilities for perception and manipulation, and is designed to run across Nvidia's Jetson edge-computing hardware.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • theaiinsider.techhttps://theaiinsider.tech/2026/09/23/nvidia-releases-isaac-ros-5-0-with-new-ai-agent-tools-for-robotics-developers/Recorded: Unknown · Article source field
  • theaiinsider.techhttps://theaiinsider.tech/2026/09/23/nvidia-releases-isaac-ros-5-0-with-new-ai-agent-tools-for-robotics-developers/Recorded: · Source history

Article published · Product

Google launches Gemini 3.8 Flash TTS models

Availability not independently checked

Google DeepMind has released Gemini 3.8 Flash TTS and Flash-Lite TTS, new text-to-speech models available in Google AI Studio and via the Gemini API, featuring custom voice creation and SynthID watermarking.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • deepmind.googlehttps://deepmind.google/blog/say-hello-to-gemini-38-text-to-speech/Recorded: Unknown · Article source field
  • deepmind.googlehttps://deepmind.google/blog/say-hello-to-gemini-38-text-to-speech/Recorded: · Source history

Article published · Security

GitHub Copilot app adds local sandboxing for AI sessions

Availability not independently checked

GitHub has introduced a local sandboxing feature in the Copilot app, allowing developers to restrict AI agent access to files, network resources, and credentials on their machines to mitigate the risks of unintended commands.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “adds”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • github.bloghttps://github.blog/changelog/2026-09-23-local-sandboxing-in-the-github-copilot-appRecorded: Unknown · Article source field
  • github.bloghttps://github.blog/changelog/2026-09-23-local-sandboxing-in-the-github-copilot-appRecorded: · Source history

Article published · Product

OpenAI launches GPT-6 Sol and Luna with 50% lower API costs

Availability not independently checked

OpenAI has released GPT-6 Sol and Luna, mid-tier models in the GPT-6 family that offer improved coding, factuality, and computer-use capabilities at significantly reduced API prices compared to their GPT-5.6 predecessors.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • openai.comhttps://openai.com/index/introducing-gpt-6-sol-and-lunaRecorded: Unknown · Article source field
  • openai.comhttps://openai.com/index/introducing-gpt-6-sol-and-lunaRecorded: · Source history

Article published · Product

SpaceXAI releases Grok 4.7 with focus on coding and cybersecurity

Availability not independently checked

SpaceXAI has launched Grok 4.7, a new flagship model featuring a larger base and extended reinforcement learning, maintaining the pricing of its predecessor.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (4)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • marktechpost.comhttps://www.marktechpost.com/2026/09/21/spacexai-releases-grok-4-7/amp/Recorded: Unknown · Article source field
  • eu.36kr.comhttps://eu.36kr.com/en/p/3993827296148482Recorded: · Source history
  • english.dainikjagranmpcg.comhttps://english.dainikjagranmpcg.com/technology/grok-47-unveiled-with-faster-coding-and-lower-ai-costs/article-32395Recorded: · Source history
  • marktechpost.comhttps://www.marktechpost.com/2026/09/21/spacexai-releases-grok-4-7/amp/Recorded: · Source history

Article published · Product

SpaceX launches Grok 4.7 with new safety and agent features

Availability not independently checked

SpaceX introduced Grok 4.7, its latest large language model, featuring improved cost-efficiency, multi-agent capabilities, and enhanced safety benchmarks for malicious request blocking.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • siliconangle.comhttps://siliconangle.com/2026/09/21/spacex-launches-grok-4-7-with-long-horizon-processing-safety-upgrades/Recorded: Unknown · Article source field
  • siliconangle.comhttps://siliconangle.com/2026/09/21/spacex-launches-grok-4-7-with-long-horizon-processing-safety-upgrades/Recorded: · Source history

Article published · Security

Zhipu AI introduces non-preservation data mechanism for model API calls

Availability not independently checked

Zhipu AI announced a forthcoming 'data content not stored' feature for its MaaS platform, designed to prevent the static storage of input and output data after model inference.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “introduces”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • news.aibase.comhttps://news.aibase.com/news/31208Recorded: Unknown · Article source field
  • news.aibase.comhttps://news.aibase.com/news/31208Recorded: · Source history

Article published · Product

Salesforce launches Koa, a vertical AI model developed with NVIDIA

Availability not independently checked

Salesforce has introduced Koa, a vertical AI model built on NVIDIA's Nemotron, designed to handle specific CRM workflows while reducing reliance on third-party providers.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • eu.36kr.comhttps://eu.36kr.com/en/p/3992455185972230Recorded: Unknown · Article source field
  • eu.36kr.comhttps://eu.36kr.com/en/p/3992455185972230Recorded: · Source history

Article published · Product

Alibaba open-sources Qwen-Image-2.1 for integrated image generation and editing

Availability not independently checked

Alibaba's Qwen team has released Qwen-Image-2.1, an open-source 7B parameter model that integrates text-to-image generation with advanced editing capabilities, including native transparent background support and multi-reference image processing.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “open”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • eu.36kr.comhttps://eu.36kr.com/en/p/3991785951198217Recorded: Unknown · Article source field
  • eu.36kr.comhttps://eu.36kr.com/en/p/3991785951198217Recorded: · Source history

Article published · Innovation

Google releases open-source EnvHarness for adaptive AI agent training

Availability not independently checked

Google Cloud AI Research has released EnvHarness, an Apache 2.0 licensed framework that dynamically modifies static training environments to target specific AI agent weaknesses, improving performance across five benchmarks.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • venturebeat.comhttps://venturebeat.com/orchestration/googles-open-source-envharness-lets-ai-agents-train-against-environments-that-evolve-with-themRecorded: Unknown · Article source field
  • venturebeat.comhttps://venturebeat.com/orchestration/googles-open-source-envharness-lets-ai-agents-train-against-environments-that-evolve-with-themRecorded: · Source history

Article published · Industry

Anthropic considers accelerating new model release ahead of IPO

Availability not independently checked

Anthropic is reportedly weighing an earlier release of a new AI model to counter the rapid enterprise adoption of OpenAI's GPT-6 Astra as it prepares for a November IPO.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “release”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • news.sbs.co.krhttps://news.sbs.co.kr/english/article.do?news_id=N1008763100Recorded: Unknown · Article source field
  • news.sbs.co.krhttps://news.sbs.co.kr/english/article.do?news_id=N1008763100Recorded: · Source history

Article published · Product

HTC launches Vive Eagle AI smart glasses in Australia

Availability not independently checked

HTC has expanded the availability of its Vive Eagle AI smart glasses to the Australian market, offering a camera-equipped wearable that integrates with Google Gemini and OpenAI’s ChatGPT.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • channelnews.com.auhttps://www.channelnews.com.au/htc-vive-eagle-ai-smart-glasses-launch-in-australia/Recorded: Unknown · Article source field
  • channelnews.com.auhttps://www.channelnews.com.au/htc-vive-eagle-ai-smart-glasses-launch-in-australia/Recorded: · Source history

Article published · Product

Alibaba Qwen Team Releases Qwen3.8-LiveTranslate for Real-Time Interpretation

Availability not independently checked

Alibaba's Qwen Team has released Qwen3.8-LiveTranslate, a real-time simultaneous interpretation model that reduces average lag to 2.3 seconds across 60 languages.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “Releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • marktechpost.comhttps://www.marktechpost.com/2026/09/19/alibaba-qwen-team-releases-qwen3-8-livetranslate/amp/Recorded: Unknown · Article source field
  • marktechpost.comhttps://www.marktechpost.com/2026/09/19/alibaba-qwen-team-releases-qwen3-8-livetranslate/amp/Recorded: · Source history

Article published · Innovation

Alibaba's Damo Academy releases open-source medical imaging model

Availability not independently checked

Alibaba's Damo Academy has released Damo Radar, an open-source vision-language model capable of identifying nearly 150 abdominal diseases from CT scans.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • newsbytesapp.comhttps://www.newsbytesapp.com/news/science/alibaba-s-ai-model-can-detect-nearly-150-abdominal-diseases/storyRecorded: Unknown · Article source field
  • newsbytesapp.comhttps://www.newsbytesapp.com/news/science/alibaba-s-ai-model-can-detect-nearly-150-abdominal-diseases/storyRecorded: · Source history

Article published · Product

Meta Launches Muse for Mac: A Personal AI Agent

Availability not independently checked

Meta has released Muse for Mac, a personal AI agent that works across your files, mail, messages, calendar, and notes.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “Launches”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Upcoming; not scheduled
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • marktechpost.comhttps://www.marktechpost.com/2026/09/19/meta-launches-muse-for-mac/Recorded: Unknown · Article source field
  • marktechpost.comhttps://www.marktechpost.com/2026/09/19/meta-launches-muse-for-mac/Recorded: · Source history

Article published · Product

PrismML releases Bonsai 2, a compressed 27B model for local devices

Availability not independently checked

PrismML has released Bonsai 2, a 5.9GB compressed version of Alibaba's Qwen3.8 27B model that runs locally on smartphones and PCs while retaining 98% of the original's benchmark performance.

Above: AI Understanding’s article text, not a quotation from a linked source. Automatic title match: “releases”. Announcement date: unknown. Article updated: .

7-day suggested check
· Due; not completed
30-day suggested check
· Due; not completed
Recorded source links (2)

Source types are unclassified. Links have not been opened by this tracker; source publication and access dates are unknown. A recorded date only indicates when the archive added the link.

  • techjuice.pkhttps://www.techjuice.pk/prismml-bonsai-2-27b-llm-compression-smartphone-pc-ai-device/Recorded: Unknown · Article source field
  • techjuice.pkhttps://www.techjuice.pk/prismml-bonsai-2-27b-llm-compression-smartphone-pc-ai-device/Recorded: · Source history

Open method · 1.0.0

Methodology & limits

01Selection

Title must contain release wording and either product wording or have the Product category. A watchlist name must appear in the title or excerpt. Mentions are not verified product ownership. This is an automatic candidate filter, not independent verification or an importance ranking.

02Unit

One canonical published article matching the watchlist and title rules; not one unique product or launch.

03Follow-ups

Suggested checks at 7 and 30 UTC calendar days after article publication (a proxy, not the announcement date). Due means the suggested date has arrived, not that a check ran. No reminders or checks are scheduled. Updated articles retain their original publication anchor.

04Coverage

Only articles returned by getAllNewsArticles, including its existing publication, suppression and media filters. That reader can silently fall back to bundled articles when the database is unavailable. Feed completeness cannot be established here; this is not a census of AI releases.

05Reproducibility

Download and retain the JSON snapshot: article text and evidence links may change with editorial updates. generatedAt is dataset assembly time, not a source access or verification time. Filters affect records, not the full-watchlist summary. Records sort newest first, undated last, then by canonical URL.

06Data source

Data source: AI Understanding’s published article archive (database and bundled articles). The existing reader does not expose which origin supplied each record or whether fallback occurred. Linked evidence is unclassified, not automatically primary.

  • Keyword selection can include rumors, commentary, comparisons and non-launch announcements, and can miss releases using other wording or names.
  • Several articles can concern one release; one article can concern several products. Mentioned organizations can be competitors rather than the releasing organization.
  • Region, account tier, waitlist, rollout stage, price and actual access remain unknown. This tracker does not measure model quality or run workflow tests.

What would count as confirmed? A separately documented human access check with time, region, account tier, product version, steps and evidence of a successful representative action. It would confirm only those conditions. This version records none; source claims, including first-party announcements, never become confirmation automatically.

The JSON contains the selection patterns, full watchlist, matched wording, dates, source provenance, unchecked status and suggested follow-ups. Missing dates remain null. Report a mismatch or missing source.

Protocol only · No results yet

Can three tools turn a source packet into usable nonprofit work?

Our first repeatable test design compares 3 tools × 10 tasks × 3 repetitions = 90 planned runs. It covers factual extraction, briefings, arithmetic, outreach, planning, privacy and uncertainty using the same synthetic source packet.

  1. Before running: select three accessible tools, freeze their versions and settings, publish the task packet and answer key, verify current prices, and approve a total cash and human-time cap.
  2. Run each task in a fresh session three times with identical inputs and rotated tool order. Preserve failures, refusals, outputs, time and cost. Stop at the approved cap.
  3. A human scores correctness, completeness, instruction compliance and usability from 0–2 each. A pass needs at least 7/8, full correctness and compliance, and no critical error. A second reviewer checks every proposed pass and critical error.
  4. Publish all denominators, disagreements, missing runs and limitations. Do not report results until runs and human review exist. Three repetitions cannot establish broad model superiority.

The full execution protocol, ten tasks, shared fixture, answer key, scoring and budget preflight live in the repository document docs/AIU-ORIGINAL-TEST-PROTOCOL.md. Tool selection, funding and execution are still pending.

Get the weekly AI Understanding briefing