Six Dimensions. One Directional Framework.
Each platform is scored 0–10 across six dimensions. Scores are directional editorial judgments based on public product documentation, pricing pages, and user-feedback themes — not validated benchmarks.
Time to Value / Vibe Creation Value
How quickly can a user describe an idea in natural language and get a usable app, workflow, UI, or feature?
QA / Assurance Before Done
Does the tool test, inspect, debug, validate, and explain issues before claiming the app is complete?
Cost Predictability
How predictable is total cost? We distinguish subscription, usage credits/tokens, hosting and database charges, third-party services, rework and maintenance, and spending limits/visibility. A predictable subscription does not guarantee predictable total cost.
True AI Agent Capability
Does the platform actually plan, edit files, build features, debug, create pull requests, deploy, or manage tasks — rather than only suggesting code?
Enterprise Readiness
Can the output run in a real enterprise environment (AWS, Azure, GCP, GitHub, Vercel, Netlify, private cloud) with identity, security, secrets, audit, and rollback?
Legacy Modernization Bonus
Can the tool help migrate, refactor, or recreate legacy applications, old codebases, spreadsheets, and outdated workflows into a modern architecture?
Scoring bands
- 9–10Strong, well-evidenced capability
- 7–8Capable with normal caveats
- 5–6Marginal / mixed evidence
- 0–4Weak or insufficient evidence
How matrix coordinates relate to scores
The X-axis (Time to Value) and Y-axis (Assurance to Production) are editorial composite positions, not a mechanical average of the six dimension scores. Quadrant labels are DERIVED from coordinates using a single threshold: a value ≥ 5.0 on an axis places a platform in the upper/right half. This is a separately assessed editorial judgment, not a mathematical derivation from the dimension scores — we do not imply it was computed from the scores. The rule guarantees coordinates and labels never contradict.
Evidence types
- Vendor-documentedCapabilities described in the vendor's own docs/changelog/pricing.
- User-reportedPublic feedback on review platforms (G2, Trustpilot, etc.) and public community discussion.
- Independently testedHands-on findings. None claimed this cycle unless explicitly stated.
We do not label documentation research as hands-on testing. We do not repeat vendor claims such as "production-ready" as independently established facts. Public-review sentiment is kept separate from capability scores and does not automatically move matrix positions.
What changed (update history)
Previous edition: Previous edition — publication date not recorded. The prior Bolt.new entry was labeled "Visionary" while plotted at (8.4, 5.8) in the Leaders area; this edition derives labels from coordinates to resolve that mismatch.
- CorrectionQuadrant labels are now DERIVED from coordinates via a single threshold (>=5.0 = upper/right). This resolves the prior Bolt mismatch: Bolt (8.4, 5.8) was labeled "Visionary" but plotted in the Leaders area; it is now consistently labeled "Leader" with a marginal-assurance caveat.
- Added platformsAdded candidate profiles: OpenAI Codex, Google Antigravity, Google AI Studio (Build mode), Emergent, Kiro, Retool (AppGen), Microsoft Power Apps vibe, and Firebase Studio. None are scored yet — listed beside the chart pending sufficient evidence.
- Lifecycle labelsEach platform now carries a lifecycle label (GA / Preview/Beta / Deprecated). Firebase Studio is marked Deprecated (sunsetting March 22, 2027); Power Apps vibe is Preview.
- Builder experienceAdded a public-feedback section per platform with source-specific ratings, review counts, dates checked, sentiment, and confidence. Sentiment is kept separate from capability scores.
- Updated capabilitiesv0 now reflects full-stack/repository workflows, CI checks, and branch protections (previously front-end-only emphasis). Cursor adds Rollouts, Security Review, Projects, and self-hosted execution. Base44 production pack dated June 16, 2026 (not a July–October release). Windsurf/Devin Desktop lineage clarified.
- MethodologyCost dimension renamed to "Cost Predictability" and now distinguishes subscription, usage credits, hosting, third-party services, rework, and spending limits. Source register added with per-claim traceability.
- PendingGoogle AI Studio Build mode source not retrieved this cycle (pending review). New candidate scoring deferred until sufficient public evidence exists.
This is a directional editorial assessment. It is not Gartner research, not a paid ranking, and not a validated benchmark. Placement is based on public product documentation, pricing pages, user-feedback themes, and directional analysis.