Sign in

Vibe Code Bench 1-100

As of Nov 3, 2026, Vibe Code Bench 1-100 (v1.0) is maintained by Vals AI. Metric accuracy (%). Ceiling 100. Saturation unknown. Contamination risk low.

Maintainer
Vals AI
Metric
accuracy (%)
Ceiling
100
Saturation
unknown
Contamination
low
Cadence
Declared cadence: as Vals updates (checked daily). Catalog snapshots: 1, latest Nov 3, 2026.
Source
Vals AI · this benchmark on Vals AI
Grading
AI judge (browser agent runs the app's workflows) — Iterative changes to an existing app; a Browser Use agent checks every workflow after each iteration.

Source: Vals AI. A proprietary private test set run by Vals; scores are republished with permission. Rows are dated the day isitsota first saw them, and only the benchmark's standard harness counts.

Categories

Where it counts

Top 5

Current top scores

  1. Muse Wonder Full · 33.978 · Nov 3, 2026
  2. Claude Opus 5.5 · 30.361 · Nov 3, 2026
  3. Claude Opus 5 · 28.533 · Nov 3, 2026
  4. Claude Fable 5.1 · 28 · Nov 3, 2026
  5. GPT-6 Astra · 27.644 · Nov 3, 2026
ModelScoreDateKind
Muse Wonder Full33.9782026-11-03snapshot
Claude Opus 5.530.3612026-11-03snapshot
Claude Opus 528.5332026-11-03snapshot
Claude Fable 5.1282026-11-03snapshot
GPT-6 Astra27.6442026-11-03snapshot
Muse Wonder26.552026-11-03snapshot
GPT-5.6 Luna22.5942026-11-03snapshot
Gemini 4 Argon21.22026-11-03snapshot
Muse Spark 1.320.4612026-11-03snapshot
Claude Haiku 5.520.4062026-11-03snapshot
GPT-5.6 Sol20.0112026-11-03snapshot
GLM 5.319.9942026-11-03snapshot
Gemini 3.8 Flash18.7722026-11-03snapshot
Kimi K318.2442026-11-03snapshot
DeepSeek V4 Pro 081317.5332026-11-03snapshot
DeepSeek V4.1 Flash16.3782026-11-03snapshot
GLM 5.3 Flash16.0332026-11-03snapshot
GPT-5.6 Terra14.8222026-11-03snapshot
Grok 4.614.752026-11-03snapshot
Mistral Large 414.2612026-11-03snapshot
Claude Sonnet 513.8222026-11-03snapshot
Qwen3 8 Max12.8332026-11-03snapshot
Minimax M39.2222026-11-03snapshot
Inkling7.3332026-11-03snapshot
Gemini 3.1 Pro Preview6.6942026-11-03snapshot