Releases

What changed on VibeToolRank, by ISO week (Monday–Sunday, Australia/Sydney). Sources: previous src/data/releases.json (git add dates in this checkout are one sanitized import) + tool JSON testedDate/lastChecked + explicit prior-batch and supplemental sentences in testSummary + docs/tests/QUEUE.md. Generated . This feed powers the newsletter’s “This week’s releases” section.

Get the Tuesday email · New launches by product date · Markdown · RSS · JSON Feed

Week of 2026-11-02

–

Tests published 9

  • · Amp · 9.0/10 — 2026-11-06 cross-check: Amp fixed our seeded cart discount bug in 3 of 3 headless Hobby Free runs (62–136s) with a free OpenRouter model
  • · Momentic · 9.6/10 — 2026-11-06 supplemental: 3/3 seeded bug-finding; published overall unchanged (9.6).
  • · Amp · 7.4/10 — Amp fixed our seeded cart discount bug in 2 of 3 headless Hobby Free runs (57–94s) with a free OpenRouter model and never touched the test f
  • · Momentic · 9.6/10 — Adapted QA task (not T3): Mo verified login + add-to-cart on Momentic's public wizard demo shop in 3/3 sessions (~60–95s, ~$0.08–$0.12 each)
  • · Qodo · 7.5/10 — Caught 3–4 of 6 planted bugs per run (avg Outcome 5.6). Never flagged the null-access bug. Fast local CLI, $0 with Gemini free. Overall 7.5.
  • · Memdoor · 9.1/10 — Memdoor fixed our seeded cart discount bug in 3 of 3 headless runs (35–47s each) with a free OpenRouter model, never touched the test file,
  • · OpenHands · 9.1/10 — OpenHands CLI headless fixed our seeded cart discount bug in 3 of 3 runs (42–51s each) with a free Gemini key, never touched the test file.
  • · TRAE · 9.0/10 — trae-agent OSS CLI fixed our seeded cart discount bug in 3 of 3 runs (19–42s each) with a free Gemini key, never touched the test file. Cost
  • · Kiro · 9.7/10 — Three Free CLI T3 runs on 2026-11-02: each fixed the seeded cart.js discount bug so node --test passed 4/4 in 17–18s, leaving cart.test.js u

Tools added 8

Page refreshes 9

  • · Bolt.new · 8.7/10 — Pricing/page re-checked 2026-11-06
  • · Cursor · 8.4/10 — Pricing/page re-checked 2026-11-06
  • · Lovable · 8.0/10 — Pricing/page re-checked 2026-11-06
  • · Momentic · 9.6/10 — Pricing/page re-checked 2026-11-06
  • · Bolt.new · 8.7/10 — Pricing/page re-checked 2026-11-05
  • · CodeRabbit · 8.1/10 — Pricing/page re-checked 2026-11-05
  • · Roomote (formerly Roo Code) — Pricing/page re-checked 2026-11-05
  • · GitHub Copilot · 8.0/10 — Pricing/page re-checked 2026-11-02
  • · OpenAI Codex · 7.4/10 — Pricing/page re-checked 2026-11-02

Week of 2026-10-26

–

Tests published 1

  • · Claude Code · 9.3/10 — Anthropic Sonnet 5.5 edition (headline): Claude Code CLI fixed our seeded cart bug in 3 of 3 runs (5.7–7.1s, 5 turns each). Test file untouc

Tools added 2

Page refreshes 4

  • · Lovable · 8.0/10 — Pricing/page re-checked 2026-10-31
  • · Netlify · 8.7/10 — Pricing/page re-checked 2026-10-31
  • · Cursor · 8.4/10 — Pricing/page re-checked 2026-10-30
  • · Railway · 9.0/10 — Pricing/page re-checked 2026-10-30

Week of 2026-10-12

–
Catalog baseline. 52 tool pages were imported on 2026-10-16 when the public git history started. Individual adds from that import are collapsed here; discovery adds and tests from the same week are listed below.

Tests published 44

  • · Reason · 8.6/10 — 3/3 runs fixed our seeded cart bug (4/4 tests pass) in 20-36s of agent time with a free OpenRouter model, $0. On the Free plan the agent's c
  • · Gemini CLI · 6.1/10 — Google-account login for individuals no longer works (the CLI says to migrate to Antigravity), so we used a free-tier API key. Run 1 (16 Oct
  • · GitHub Copilot CLI · 9.2/10 — 3/3 runs fixed our seeded cart bug (4/4 tests pass) in 22-26s on Copilot Free, using 0.16-0.27 AI Credits each. It used our existing gh logi
  • · Harness (Autonomous) · 8.4/10 — Harness started OpenCode 3 times on our seeded cart bug and each run fixed it (4/4 tests pass) in 19-38s, at $0 with a free model. Before th
  • · Skisso · 8.3/10 — On our T1 todo app, Skisso turned a visual edit (Add button radius 10px → 20px, label "Add task") into an exact, scoped handoff with selecto
  • · Zed · 6.8/10 — We ran Zed's Agent on the free Personal plan with our own OpenRouter key pointed at the free model cohere/north-mini-code:free (Pro trial wa
  • · Aider · 9.3/10 — 3/3 runs fixed our seeded cart bug (4/4 tests pass) in 9-13s with a free OpenRouter model, without touching the test file. Install needed uv
  • · Base44 · 8.2/10 — Our T1 todo prompt gave a working preview in about 3.4 minutes. Add, complete, delete, All/Active/Done filters and reload persistence all pa
  • · Better Auth · 8.2/10 — Email/password sign-up, sign-in, session, wrong-password rejection, 3/3 runs passed; needed node:sqlite and the new `npx auth migrate` CLI.
  • · Bolt.new · 8.7/10 — Our T1 todo prompt gave a working preview in about 68s. Add, complete, delete, filters and localStorage persistence all passed with zero fix
  • · Builder.io · 6.7/10 — Our T1 todo prompt took about 12-13 minutes to a working preview, partly because of a blank loading screen that needed reopening. Add, compl
  • · Clerk · 8.1/10 — 3/3 runs: sign-up (Backend API createUser), password sign-in plus new-device email code, active session check, wrong password rejected (422)
  • · Cline · 9.2/10 — 3/3 headless CLI runs fixed our seeded cart bug (4/4 tests pass) in 24-35s with a free OpenRouter model, without editing the test file.
  • · Cloudflare Pages & Workers · 8.4/10 — 2/3 runs: static page + Pages Function deployed from local files; live in 10-19s, redeploy live in 6-8s, projects deleted cleanly. Run 2 fai
  • · CodeRabbit · 8.1/10 — On our seeded R1 pull request with 6 planted bugs, CodeRabbit posted a summary in 14s and a full review in about 2.5 minutes. It caught the
  • · Context7 · 9.1/10 — 3/3 runs: resolved 'astro' to /withastro/docs and returned current Astro 5 content-collection docs (defineCollection, glob loader, src/conte
  • · Convex · 8.7/10 — 3/3 runs: push functions, insert and query rows, arg validation, all OK in about 7-8s each (warm). The self-hosted Docker images on ghcr.io
  • · Cursor · 8.4/10 — In one agent turn with no follow-ups, Cursor built our T1 todo app as a single index.html in about 2.5 minutes and tested it in its built-in
  • · Devin Desktop (formerly Windsurf) · 8.1/10 — Devin Desktop, Cognition's renamed Windsurf editor, built our T1 todo app in about 1 minute with no follow-ups and previewed it after one 'F
  • · Figma Make · 8.0/10 — Our T1 todo prompt gave a working preview in about 3-4.5 minutes, after a transient 'Live preview failed' message that resolved itself. Add,
  • · GitHub Copilot · 8.0/10 — Our first chat (Auto routed to GPT-5 mini) failed 3 times with 'Sorry, no response was returned'. A new chat in Copilot mode, where Auto pic
  • · Google AI Studio · 7.5/10 — Our T1 todo prompt gave a usable preview in about 3 minutes. All features passed, including reload persistence. It heavily over-delivered fe
  • · Greptile · 7.8/10 — On our seeded R1 pull request with 6 planted bugs, Greptile reviewed about 2.5 minutes after we asked with '@greptileai review'; it didn't a
  • · Jules · 8.0/10 — Jules is an async coding agent: it delivered our T1 todo app as a GitHub pull request after about 14 minutes, about 5 of which it appeared s
  • · Lovable · 8.0/10 — Our T1 todo prompt gave a working preview in about 90s and used 0.70 credits. In the editor preview, add, complete, delete, filters and relo
  • · Neon · 9.5/10 — 3/3 runs: one API call created a Postgres branch with compute that answered queries ~2s later; created a todos table, inserted, updated and
  • · Netlify · 8.7/10 — 3/3 runs: static page + serverless function deployed via one CLI command, live 200 in ~17-20s. New sites defaulted to Netlify-login protecti
  • · OpenAI Codex · 7.4/10 — Using a custom OpenRouter provider and a small free model, Codex CLI fixed our seeded cart bug in 2 of 3 runs (19-76s). In the failed run th
  • · OpenCode · 9.2/10 — 3/3 runs: fixed a seeded cart-discount bug (2 failing tests) in 18-22s with a free Zen model and no account or API key, without touching the
  • · Paddle · 8.1/10 — 3/3 sandbox runs: created a $9/month product and price, then a transaction with a checkout link, in under 1s per run. It needed a default pa
  • · Playwright MCP · 8.8/10 — 3/3 runs: MCP server started via npx, navigated to a page, returned an accessibility snapshot and clicked a link by ref in ~3s total. Functi
  • · Polar · 8.8/10 — 3/3 sandbox runs: a $9/month product was created in one call, a checkout session returned a hosted page that loaded (HTTP 200), and the prod
  • · PostHog · 9.0/10 — 3/3 runs: an event sent with one HTTP POST was queryable with HogQL 11-18s later, with the right properties. Query results are cached by def
  • · Railway · 9.0/10 — 3/3 runs: Node app uploaded from local files (no GitHub) via railway up, public URL live in 30-51s, redeploy live in 27-48s. Public domain n
  • · Render · 8.9/10 — 3/3 runs: public Docker image deployed as a free web service with one API call, live 200 in ~60s each. Resume from suspended took 11.5s to f
  • · Replit Agent · 7.6/10 — Our T1 todo prompt gave a working preview in about 6-9 minutes. Replit built an Expo mobile app shown in a web phone preview rather than a p
  • · Semgrep · 8.1/10 — Seeded-repo scan, 3 runs: 6/10 issues found, 0 false positives. Missed secrets, MD5, yaml.load, DOM XSS.
  • · Sentry · 9.3/10 — 3/3 runs: a Node script captured an exception with a tag and release, and the event, including its stack trace, was queryable in Sentry's AP
  • · Snyk · 5.6/10 — Snyk is a SAST and security scanner, not a general AI reviewer. On our seeded R1 pull request it commented within about 9s and correctly fla
  • · Sourcery · 8.8/10 — On our seeded R1 pull request with 6 planted bugs, Sourcery caught all 6 with sensible severities, including the SQL injection and, through
  • · Stripe · 9.2/10 — 3/3 runs in test mode: product+price, Checkout Session, PaymentIntent confirmed with pm_card_visa, full refund, 7 events verified via Events
  • · Supabase · 9.2/10 — Cloud: project provisioned in ~9s, table + REST write/read + email signup/login/wrong-password rejected passed 3/3 runs (~12s task). Email c
  • · v0 · 8.3/10 — Our T1 todo prompt was done in about 94s, with a preview link at 25s. v0 tested its own work in its browser, found a missing lib/utils file
  • · Vercel · 9.4/10 — 3/3 runs: static page + Node serverless function deployed from local files via vercel deploy --prod; both URLs 200 in 15-23s, redeploy live

Tools added 10

Get the weekly test results Hands-on scores, timings and free-tier gotchas, once a week. Sent Tuesday morning, Australia/Sydney time (Monday evening in the US). Free. Unsubscribe in one click from any issue. Read past issues · Privacy policy.