# Morning digest: Code Review, Gemini 4 Argon, Kling 4.0, and RTX Spark > Code Review is generally available on Team and Enterprise, Gemini 4 Argon is in Antigravity, Kling 4.0 is fully out, RTX Spark laptops start at $2,599, and Meshy Edit changes one part of a model. Published 2026-10-16T08:36:00-07:00 · Daily Digest · Canonical: https://news.littlewolflabs.com/posts/2026-10-16-morning-digest Checked October 16, 2026. Each item links the page it came from. ## Claude Code Review is generally available on Team and Enterprise [ClaudeDevs](https://x.com/ClaudeDevs/status/2110866210573386065) and Anthropic's [general-availability note](https://claude.com/resources/articles/code-review-is-now-generally-available) say Code Review is out of research preview for Team and Enterprise. Admins set Quick, Standard (the default), or Deep per repository. Anthropic's claim is that most reviews now finish about three times faster, from about 20 minutes to 6–10. If agent pull requests are already piling up, this is the one worth turning on. Set the spend limit before the first review. The [costs page](https://code.claude.com/docs/en/code-review-costs) says there is no flat per-review price. Reviews are billed on token usage through usage credits, separate from plan limits, and the charge goes to the organization, not the person who opened the pull request. The cost still shows on the Anthropic bill if the rest of Claude Code runs on Amazon Bedrock, Google Cloud's Agent Platform, or Microsoft Foundry. The org monthly spend limit is set at [claude.ai/admin-settings/usage](https://claude.ai/admin-settings/usage). The [promotional-credits page](https://support.claude.com/en/articles/17603594-free-code-review-promotional-credits-for-claude-enterprise) is more precise than the announcement line. Eligible Enterprise organizations claim once, and the amount set at claim is between $10,000 and $50,000 based on active Claude Code seats, so "at least $10,000" is the floor, not a flat grant for every org. The claim window runs from October 15, 2026 through 23:59 UTC on January 15, 2027. Credits expire 30 days after the claim. Reseller customers, GitHub Enterprise Server and .ghe.com, trials, and Team, Pro, Max, and Free plans are excluded, as is anyone who already took an earlier offer. ## Gemini 4 Argon is in Antigravity, on a weekly cap [Antigravity](https://x.com/antigravity/status/2110866531932316011) says Gemini 4 Argon is rolling out to Google AI Pro, with a separate weekly-only Argon quota on top of the Gemini 3.8 Flash bucket. The [Antigravity blog](https://antigravity.google/blog/gemini-4-argon-in-google-antigravity) and [9to5Google](https://9to5google.com/2026/10/15/gemini-4-argon-is-now-rolling-out-to-googles-antigravity-for-ai-pro-subscribers/) report the same rollout. No numeric weekly cap is published. The flagship is in the coding tool. Pace yourself. ## Kling 4.0 is fully available [Kling](https://x.com/Kling_ai/status/2110747537187709424) says Kling 4.0 is fully available, with a refreshed interface plus an MCP server and a CLI. The [user guide](https://kling.ai/quickstart/kling-4-model-user-guide) lists a single pass of 3–30 seconds, up to 10 keyframes, up to 15 references, and 720p, 1080p, and 4K. The tweet's "Coming" section is cut off in the public post, so HDR, a 2-minute extension, and a developer API "coming soon" are from [Kie.ai's write-up](https://kie.ai/blog/kling-4-0-release-analysis-30-second-clips-keyframes-pricing), not text visible in that capture. The same Kie.ai piece says testers report 15 credits per second at 1080p (Khazix measured 450 credits for a 30-second 1080p clip, which is 30 times 15) and 10 credits per second at 720p on a rate card from @Chinazhidx. That 15-credits-per-second figure is tester-reported. It is not on Kling's page. ## RTX Spark laptops go on sale from $2,599 NVIDIA's [October 16 announcement](https://nvidianews.nvidia.com/news/nvidia-rtx-spark-pcs-available-today-a-new-beginning-for-windows-pcs) says RTX Spark PCs are available today from ASUS, Dell, HP, Lenovo, Microsoft, and MSI, starting at $2,599, with up to 128GB of unified memory. Named machines include the ASUS ProArt P14 and P16, Dell XPS 16, HP OmniBook Ultra and OmniBook X 14, Lenovo Yoga Pro 9n, Microsoft Surface Laptop Ultra, and MSI Prestige N16 Flip AI+ and Prestige N16. NVIDIA also says the lineup reaches up to 1 petaflop of FP4. This is the first thin Windows laptop aimed at big local models. The open question is whether it beats a desktop GPU for Blender and Unreal. In a separate [October 15 note](https://nvidianews.nvidia.com/news/adobe-and-nvidia-expand-collaboration-to-speed-agentic-creation), Adobe says Photoshop, Premiere, and Substance 3D optimized for RTX Spark deliver up to 2x faster AI processing, editing, color, and effects. That speed line is a vendor claim. ## Meshy Edit changes the part you select [Meshy](https://x.com/MeshyAI/status/2110989727302381915) says you select a part, describe the change, and generate a new model while the rest stays put. The [October 16 release](https://www.prnewswire.com/news-releases/meshy-debuts-new-in-house-ai-model-with-meshy-edit-letting-users-change-selected-parts-of-a-3d-model-with-one-prompt-302911259.html) puts that at about one minute and 20 credits per edit, on high-poly Meshy 6 and Meshy 7.1, including Ultra 2K, for all registered users. Textured models switch to untextured for the edit, you can add up to five reference images, and the original is kept. Ultra 4K, low-poly, and uploads are not supported yet. The point is a part swap without a full regeneration or a hand edit in Blender. ## Muse Wonder is in Cursor at the same price as Opus 5.5 [Cursor](https://x.com/cursor_ai/status/2110867556693430293) lists Muse Wonder at $4 per million input tokens and $20 per million output tokens, with 46.1% on CursorBench at default effort, the highest Muse score on that bench. CursorBench is Cursor's own benchmark. [Claude's pricing](https://platform.claude.com/docs/en/about-claude/pricing) lists Opus 5.5 standard at the same $4 and $20. Test Wonder against Opus 5.5. Do not auto-switch. ## 404 Media reports Accenture will fund $29 of Claude a month [404 Media](https://www.404media.co/accenture-tells-employees-they-can-now-spend-only-29-on-claude-per-month/) reports that an internal Accenture memo caps company-funded Claude at $29 a person a month, effective October 1. The report says overage is billed to projects, and that some staff had been spending up to $1,000 a month. This is 404 Media's account of an internal announcement, not a page on Accenture's site. It is the token-cost crunch in one number. ## ChatGPT ads can target a list of emails and phone numbers OpenAI's [custom audiences help](https://help.openai.com/en/articles/20001346-set-up-custom-audiences-for-your-campaign) says advertisers upload email or phone lists, raw or SHA-256 hashed, plus a Google Advertising ID. Inclusion and bid adjustments need at least 25,000 matched users. OpenAI recommends 100,000. A smaller audience can only exclude people. Bid multipliers run from 0.1x to 10x, and the highest matching multiplier wins. Ads Manager does not show individual matched users. Uploaded files are deleted after processing, usually within 24 hours. For the EEA and Switzerland, the page says to avoid custom audiences while ads personalization is unavailable there. The [developer docs](https://developers.openai.com/ads/custom-audiences) use the same 25,000 figure as the planning threshold. [Ads in ChatGPT](https://help.openai.com/en/articles/20001047-ads-in-chatgpt) says ads may appear on Free and Go. Plus, Pro, Business, Enterprise, and Edu do not have ads. Turning personalization off still lets the current chat shape ads. It does not use other chats, ads history, or topics. [The Information](https://www.theinformation.com/articles/openais-accelerates-ad-sprint-meta-hires-slashed-prices) reports about 60,000 advertisers and lower prices. [404 Media](https://www.404media.co/chatgpt-ads-target-your-phone-number-email/) reported the phone-and-email targeting. Paid plans stay ad-free. ## Short items Users say Amazon is blocking Grok Bots from placing orders. That is a user claim. Amazon has not posted a statement. Elon Musk replied, ["Maybe we should start a competitor to Amazon."](https://x.com/elonmusk/status/2110715556328579219) [Bloomberg Government](https://news.bgov.com/bloomberg-government-news/senate-preparing-to-roll-out-anthropics-claude-for-staff-use) reports the Senate is poised to give staff Claude access, according to three people briefed. It has not been announced. ## Video ideas Can a $2,599 RTX Spark laptop replace a desktop for local AI plus Blender/Unreal? Meshy Edit into Blender/UE5: swap parts on a generated character, drop into a Gaea landscape. Kling 4.0 30-second shots: what a usable clip really costs after rerolls.