Antigravity extra reset complete
After the Neon cache change, Antigravity has reset every user's Gemini quota; the same quota can get about 22% more work done.
🔄 RESETS · CARDS · TEASERS · NEWS
Extra quota and reset cards from AI subscriptions, logged by date — plus card expiry dates. New models, open-source releases, free access and plan changes get a note too; normal weekly/monthly resets are not included.
Click a type to filter, then All to reset.
Page 1 of 7 · 25 on this page
Older records are on earlier pages.
After the Neon cache change, Antigravity has reset every user's Gemini quota; the same quota can get about 22% more work done.
Varun Mohan says Antigravity is enlarging the Neon cache pool: cache hit rate goes up by roughly 4% and the same quota should get at least 20% more work done. A follow-up in the thread says workloads with longer gaps between model calls were hit hardest because the cache was evicted; a reset will follow the change.
Tencent's Hy team has open-sourced Hy4, Hy4-FP8 and the pretrained Hy4-base weights on Hugging Face. Hy4 is a new-generation multimodal Mixture-of-Experts flagship with a 770B-parameter language backbone, 49B text parameters plus 1.23B vision parameters active per token, text/image/video understanding and a 1,048,576-token context under the Tencent Hy License.
Neon is adapted from Gemini 4 Carbon for high-frequency coding, analysis and agent workflows. Google says it scores 61 on the Artificial Analysis Intelligence Index (xhigh), third overall, with API pricing at $3/$15 per million input/output tokens and a 95% cached-input discount.
Tencent's Hy team has published Hy4-base on Hugging Face, the pretrained base version of its new-generation Mixture-of-Experts flagship model, supporting text, image and video understanding with a 770B-parameter language backbone. Weights are downloadable under the custom Tencent Hy License, with mirrors on ModelScope and GitCode and a GGUF variant.
For Antigravity's first anniversary, Varun Mohan announced a quota reset for all models for all users.
Astra targets high-difficulty professional work and is priced 20% below GPT-6 Astra on the API; Sol is more efficient and aimed at everyday tasks. Plus, Pro, Enterprise, and Edu can use them the same day.
Tencent Hunyuan published Hy4 flash, FP8 and base open weights: about 220B total parameters (including 63.5B Engram), roughly 12B active per token, text and image input, and a 1M context window under the Tencent Hy License.
Tibo announced that banked resets are activated and will roll out to paid accounts; users can choose when to use them in Settings > Usage.
Covers all models and limits and can be used any time before December 10, 14:00 PT; it does not move your weekly reset day.
xAI is again sending resets to Grok Bot users.
Anthropic announced a Claude usage limit reset via @lydiahallie.
Max / Ultra / Team Premium get 1–3 Fable reset cards based on subscription length, usable until March 8, 2027; Pro and Team standard seats get a separate free week.
Anthropic releases Fable 5.2 and Mythos 5.2; Fable is cheaper and faster, and is open to all users.
Pro and Team standard seats get a free week of Fable from November 9-16, with free usage capped at 50% of the weekly limit; the offer was not extended.
The Codex banked reset issued 2026-10-08 expired on 2026-11-07; it lapses if unused and is usually not reissued.
Varun Mohan said Antigravity is resetting Gemini and Argon model quotas for everyone so users can keep building over the weekend after a week of updates.
The Codex banked reset issued 2026-10-07 expired on 2026-11-06; it lapses if unused and is usually not reissued.
The Codex banked reset issued 2026-10-08 expires in 3 days (2026-11-07); if unused, it lapses on expiry.
The Codex banked reset issued 2026-10-07 expires in 3 days (2026-11-06); if unused, it lapses on expiry.
The 28-day run ends, granting two banked resets (1/2, 2/2).
In his Day 26 update post, Tibo added: the next two days will both be resets, and they will be banked.
In his Day 23 post Tibo said "today is a reset"; he then added it would land by midnight PST and confirmed completion at 06:29 UTC.
OpenAI releases GPT-6.1 Terra, positioned between 6.1 Sol and Luna, with lower pricing and aimed at high-frequency everyday tasks.
Kimi apologized for subscription throttling and "unnecessary consumption" and gave all subscribers a one-time restore of 25% of their monthly quota, which it described as one week of usage.
Nothing matches these filters.