<?xml version="1.0" encoding="UTF-8"?>
<!-- Generated by tools/build-pages.js from changelog.js. Deterministic: lastBuildDate is the newest changelog date, never now. -->
<rss version="2.0">
<channel>
<title>Vibe-coder frontier: what changed</title>
<link>https://vibecoderfrontier.com/#changes</link>
<description>Catalog updates from the daily check, with the source each one came from.</description>
<lastBuildDate>Sun, 25 Oct 2026 00:00:00 GMT</lastBuildDate>
<item>
<title>What changed on 25 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-25</guid>
<pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;OpenCode Go cut Gemini 4 Argon usage pricing by 50% through 30 November. &lt;a href=&quot;https://opencode.ai/docs/go/&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 21 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-21</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 21 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-21</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Corrected the OpenCode Go Plus note: official monthly model caps run $60–$240, not $60–$180; hours and price are unchanged. &lt;a href=&quot;https://opencode.ai/docs/go/&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 20 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-20</guid>
<pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 20 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-20</guid>
<pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Added Meta Muse Code in all four subscription tiers — Everyday ($5/mo), High ($15), Power ($50, adds Muse Wonder Full) and Maximum ($200). Hours are converted cautiously from Meta&amp;#39;s published prompts-per-5-hours windows and tagged guessed until lived reports land. Muse has no verified row in the current coding-agent benchmark suite yet, so the plans are listed but not ranked. &lt;a href=&quot;https://dev.meta.ai/products/muse-code/&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Added the unified X Subscription coding tiers — X Plus ($8/mo), X Premium ($30), X Super ($100) and X Ultra ($200) — each bundling Grok Build and Cursor on one shared weekly pool. Hours are carried from the retired SuperGrok seats and haircut well below the official 4x/15x/50x usage banners, tagged guessed. &lt;a href=&quot;https://x.ai/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://grok.com/plans&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;SuperGrok, SuperGrok Plus and SuperGrok Heavy are now marked paused: they continue only for existing subscribers under the new X Subscription, and cannot be bought new. Cursor Pro ($30) is the same unified Premium subscription, so it is now tracked once as X Premium instead of twice. &lt;a href=&quot;https://x.ai/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://cursor.com/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Caught up on Codex plans: added Codex Go ($8/mo, GPT-6 Luna in the desktop app only) and Codex Pro 500 ($500/mo, 25x the Plus allowance plus Ultrafast, which burns allowance 8x faster), and unpaused the $200 tier, which reopened as Pro 200 at 10x Plus (the old 20x deal ends 29 Oct). The $100 tier is now sold as Pro 100. New hours are haircut guesses, not banner multiples. &lt;a href=&quot;https://learn.chatgpt.com/docs/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;MiniMax renamed its Token Plan to M Plan on 30 Sep: Plus is now M Plan Go ($22/mo), Max is M Plan Explore ($55), Ultra is M Plan Build ($132). Existing subscribers were migrated and quotas are unchanged, so hours carry over; names and prices updated. &lt;a href=&quot;https://platform.minimax.io/docs/m-plan/faq&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Added the paid Google AI tiers for Antigravity (AI Pro about $20/mo, AI Ultra $100 at 5x Pro limits, AI Ultra $200 at 20x) and the missing Copilot tiers (Pro $10/mo, Max $100, both metered in GitHub AI Credits). Hours are scaled cautiously from the existing seats, tagged guessed, and these plans are listed but not ranked until they have a verified benchmark mapping. &lt;a href=&quot;https://antigravity.google/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://github.com/features/copilot/plans&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 15 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-15</guid>
<pubDate>Thu, 15 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 14 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-14</guid>
<pubDate>Wed, 14 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 13 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-13</guid>
<pubDate>Tue, 13 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 13 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-13</guid>
<pubDate>Tue, 13 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 13 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-13</guid>
<pubDate>Tue, 13 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Cursor&amp;#39;s standard plan is now called Premium and costs $30 per month under the new unified lineup. &lt;a href=&quot;https://cursor.com/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://x.ai/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Cursor&amp;#39;s included usage now covers only Grok models and refreshes weekly; third-party models bill separately at API list price. &lt;a href=&quot;https://cursor.com/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://x.ai/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Cursor Pro+ is paused for new signups; the new lineup replaces it with Super at $100 per month. &lt;a href=&quot;https://cursor.com/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://x.ai/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Cursor Pro+ hours stay as last measured; Super&amp;#39;s advertised usage multiplier was not used to re-derive them. &lt;a href=&quot;https://cursor.com/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 10 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-10</guid>
<pubDate>Sat, 10 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 10 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-10</guid>
<pubDate>Sat, 10 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Claude Max 5x now includes $100 per month in Claude Platform API credits, which sit outside the subscription hours we track. &lt;a href=&quot;https://www.anthropic.com/pricing&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Claude Max 20x now includes $200 per month in Claude Platform API credits, which sit outside the subscription hours we track. &lt;a href=&quot;https://www.anthropic.com/pricing&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://www.reddit.com/r/ClaudeAI/comments/1x20kdf/200_free_api_credits_every_month_from_anthropic/&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 9 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-09</guid>
<pubDate>Fri, 09 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;MiniMax now sells M Plan tiers Go, Explore and Build (1x/3x/7.5x usage), so Plus appears legacy; the row stays listed while prices and hours are re-verified. &lt;a href=&quot;https://platform.minimax.io/docs/token-plan/faq&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;MiniMax now sells M Plan tiers Go, Explore and Build (1x/3x/7.5x usage), so Max appears legacy; the row stays listed while prices and hours are re-verified. &lt;a href=&quot;https://platform.minimax.io/docs/token-plan/faq&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;MiniMax now sells M Plan tiers Go, Explore and Build (1x/3x/7.5x usage), so Ultra appears legacy; the row stays listed while prices and hours are re-verified. &lt;a href=&quot;https://platform.minimax.io/docs/token-plan/faq&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 8 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-08</guid>
<pubDate>Thu, 08 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;The GLM-5.3-Flash unlimited evening campaign and the all-day off-peak window both ended on 7 Oct; GLM Max hours are unchanged. &lt;a href=&quot;https://docs.z.ai/devpack/overview&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 8 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-08</guid>
<pubDate>Thu, 08 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings refreshed from current Artificial Analysis coding-agent evaluations. Each score identifies the agent, model and effort; scores are out of 100. Plans without a verified mapping are listed but excluded from ranking. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 7 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-07</guid>
<pubDate>Wed, 07 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Supported models are now checked daily for every provider. Added MiniMax M3.1 Flash Preview with its access restriction, current Cursor and Copilot models, and the full OpenCode Go family list. Qwen and Antigravity show the models supported by the specific listed plan. &lt;a href=&quot;https://platform.minimax.io/docs/guides/text-generation&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://opencode.ai/docs/go/&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://cursor.com/docs&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://docs.github.com/en/copilot/reference/ai-models/supported-models&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://antigravity.google/docs/models&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://www.alibabacloud.com/help/en/model-studio/coding-plan&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 7 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-07</guid>
<pubDate>Wed, 07 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality rankings now use current Artificial Analysis Coding Agent Index v1.5 evaluations for Fable 5.1, Sonnet 5.5, Opus 5.5, GPT-6.1 Sol, Astra, GLM-5.3 and Grok 4.7. Scores are out of 100 with agent and effort shown. Plans without a verified mapping are excluded; estimated subscription hours are unchanged. &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents&quot;&gt;source&lt;/a&gt; &lt;a href=&quot;https://artificialanalysis.ai/agents/coding-agents/comparisons/claude-code-vs-codex&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 4 Oct 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-10-04</guid>
<pubDate>Sun, 04 Oct 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;SuperGrok Plus no longer appears on xAI&amp;#39;s plans page, so the seat is paused for new signups and kept listed for reference. &lt;a href=&quot;https://grok.com/plans&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;The paused SuperGrok Plus row now explains that xAI dropped the Plus tier from its plans lineup in early October. &lt;a href=&quot;https://grok.com/plans&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 29 Sep 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-09-29</guid>
<pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;OpenCode added a Go Plus tier at $40 per month with roughly triple Go&amp;#39;s per-five-hour request limits; we list it at a conservative double Go&amp;#39;s hours until lived reports arrive. &lt;a href=&quot;https://opencode.ai/go&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;MiniMax&amp;#39;s Token Plan now deducts by actual usage from one shared quota pool; the vendor says Plus keeps its five-hour usage plus eight percent, so the listed hours stay put for now. &lt;a href=&quot;https://platform.minimax.io/docs/token-plan/faq&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 25 Sep 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-09-25</guid>
<pubDate>Fri, 25 Sep 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Quality is now a public benchmark instead of a grade: each plan is mapped to the nearest measured model on SWE-rebench (swe-rebench.com, run by Nebius) and scored on its pass rate across your whole week, not just the best plan&amp;#39;s grade. Plans with no benchmark row (Cursor, Copilot) or only old data (Qwen) are listed but not ranked. &lt;a href=&quot;https://swe-rebench.com/&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;Stacks within a point of the best score now count as a tie, and the cheapest one is shown as #1, so spending more has to buy something.&lt;/li&gt;&lt;/ul&gt;</description>
</item>
<item>
<title>What changed on 24 Sep 2026</title>
<link>https://vibecoderfrontier.com/#changes</link>
<guid isPermaLink="false">2026-09-24</guid>
<pubDate>Thu, 24 Sep 2026 00:00:00 GMT</pubDate>
<description>&lt;ul&gt;&lt;li&gt;Added SuperGrok Heavy + Build at $300/mo (grok.com/plans); xAI publishes no usage multiplier, so its hours are tagged guessed. &lt;a href=&quot;https://grok.com/plans&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;li&gt;GLM Pro is now $80/mo (was $76) and GLM Max $168/mo (was $120), per Z.ai&amp;#39;s monthly pricing page; hours unchanged. &lt;a href=&quot;https://z.ai/subscribe&quot;&gt;source&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</description>
</item>
</channel>
</rss>
