Compare subscriptions
Check the ChatGPT Plus price, compare Claude Pro vs Max and weigh subscriptions against API costs.
Compare AI subscription pricesbuffer’s AI buying notes · Stay curious
Which AI subscription is worth it?Know the limits. Choose your hardware.
Compare Claude Pro vs Max and ChatGPT Plus, understand usage limits and resets, and find the Mac mini or GPU you need to run a local LLM.
Start with the question you have
Track usage limits
Follow Codex usage limit resets, Claude usage windows and rule changes, with links to the sources.
Open the Codex limit trackerCompare coding tools
Claude Code vs Codex: choose a coding assistant by workflow, cost and the limits you will run into.
Compare Claude Code vs Codex Claude Pro vs Max: worth the upgrade?This week’s popular comparisons
Four guides to start with
AI hardware
Mac mini for AI, or a GPU for a local LLM? Start with memory and VRAM.
Codex quota resets
Tibo’s reset announcements, on your local calendar.
Tibo @thsottiauxOctober 2026
Your time
Asia/Shanghai
Direct quota resetReset credits issued11 · this month
| Mon | Tue | Wed | Thu | Fri | Sat | Sun |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | |||
| 5 | 6 | 7 | 8 | 9 | 10 | 11 |
| 12 | 13 | 14 | 15 | 16 | 17 | 18 |
| 19 | 20 | 21 | 22 | 23 | 24 | 25 |
| 26 | 27 | 28 | 29 | 30 | 31 |
Reset details1
2026-10-03
Direct quota resetConfirmed
UTC+08:00
Global reset all paid ChatGPT 10am PST (Sol load spike). Confirm: all propagated.
Original post ↗Compare all cities
2026-10-07
Direct quota resetConfirmed
UTC+08:00
28-day Day 2 vote: community chose reset over four ships.
Original post ↗Compare all cities
2026-10-08
Reset credits issuedConfirmed
UTC+08:00
Day 3 / 40M: banked reset into every paid account.
Original post ↗Compare all cities
2026-10-11
Direct quota resetConfirmed
UTC+08:00
Day 6 confirm ‘All propagated’. Day post + EOD reply earlier.
Original post ↗Compare all cities
2026-10-13
Direct quota resetConfirmed
UTC+08:00
Day 8 confirm ‘It’s done’. Thread said global by EOD PST.
Original post ↗Compare all cities
2026-10-15
Direct quota resetConfirmed
UTC+08:00
Day 10/ Reset.
Original post ↗Compare all cities
2026-10-18
Direct quota resetConfirmed
UTC+08:00
Day 13/ Le limit reset.
Original post ↗Compare all cities
2026-10-20
Direct quota resetConfirmed
UTC+08:00
Day 15/ still cooking... and reset. Enjoy.
Original post ↗Compare all cities
2026-10-23
Direct quota resetConfirmed
UTC+08:00
Day 17 reset live. Tibo announced by 10am PT Oct 22; user sighting ~01:17 Beijing Oct 23 (17:17Z) that Codex week refreshed. No separate Tibo confirm; Free may lag. Confirmed from user reports (e.g. IndependentEco).
Original post ↗Compare all cities
2026-10-24
Direct quota resetAnnounced · unconfirmed
By UTC+08:00
Day 18 reset; lands before 10am Pacific (PDT) Oct 23 with small ship. Stored as 17:00Z = 10:00 America/Los_Angeles. Announce≠delivered.
Original post ↗Compare all cities
2026-10-25
Direct quota resetConfirmed
UTC+08:00
Day 20 / Rest and reset live. Announced by midnight PST; Tibo confirmed "It is all propagated" at https://x.com/thsottiaux/status/2114237083652882932 (06:06Z = 14:06 Beijing Oct 25). Plans not specified.
Original post ↗Compare all cities
- Shanghai
- Berlin
- New York
- Tokyo
- Madrid
- Mexico City
- San Francisco · OpenAI
Dates follow the selected city and can cross day or month boundaries. “By” marks a deadline; announcements do not confirm completion. No fixed reset schedule.
Latest entries
- Direct quota reset · Confirmed
Reset time not specified · Read the original post ↗
Tibo announced a reset (Day 21) at 13:21 Beijing time on Oct 26 (22:21 PDT Oct 25), without an exact landing time. Third-party trackers logged it landing around 14:10–15:00 Beijing time on Oct 26, with the first user reports around 14:00.
- Direct quota reset · Confirmed
Reset time · Read the original post ↗
Day 20 / Rest and reset live. Announced by midnight PST; Tibo confirmed "It is all propagated" at https://x.com/thsottiaux/status/2114237083652882932 (06:06Z = 14:06 Beijing Oct 25). Plans not specified.
- Direct quota reset · Confirmed
Reset time · Read the original post ↗
Day 17 reset live. Tibo announced by 10am PT Oct 22; user sighting ~01:17 Beijing Oct 23 (17:17Z) that Codex week refreshed. No separate Tibo confirm; Free may lag. Confirmed from user reports (e.g. IndependentEco).
Notes & observations
All articles Recent notes
Ollama llama.cpp Errors: A Quick Lookup of Exact Messages and Fixes (2026)AI
A quick lookup of Ollama and llama.cpp error messages, grouped by pulling, loading and starting. Each row gives the exact text, the likely cause, whether we tested it, and a link to the full fix.
Read article
OpenAI Codex Error Codes: A Quick Lookup for API, ChatGPT and Codex Messages (2026)AI
A quick lookup of OpenAI API, ChatGPT and Codex error messages. Each row gives the exact text, the likely cause in one line, where that answer…
Read article
Claude Code auto mode temporarily unavailable: What It Means and 5 Easy Fixes (Tested)Claude
Claude Code auto mode temporarily unavailable means the classifier model that checks actions in auto mode did not answer, so the action was skipped. We reproduced…
Read article
Ollama “Downloaded in a Format That Is No Longer Supported”: A Confusing Error, Clearly FixedAI
Since Ollama 0.40.4, /v1/systemone refuses decision models in a format that is no longer supported. Official models like nimble, tev1 and clef just need a fresh…
Read article
Claude Code monthly spend limit: Fix “You’ve hit your monthly spend limit” Fast (Tested)Claude
Claude Code monthly spend limit means your plan usage ran out and your usage credits hit their monthly cap. We tested the message family on Claude…
Read article
Ollama “pull model manifest: 412”: The Simple Fix, and Why Upgrading Is Not Always EnoughAI
Ollama shows "Error: pull model manifest: 412" when the model you are pulling needs a newer Ollama than the one you run. Upgrade the official way,…
Read article
Ollama “unknown model architecture”: The Frustrating Error Explained, With Real Fixes (2026)AI
Ollama shows "error loading model: unknown model architecture" when the model file uses an architecture its bundled llama.cpp does not know yet. Upgrade Ollama, use the…
Read article
About me
I’m buffer.
I talk AI and hardware on X. When a conversation needs more room, I continue it here.