Sign in

CyberBench

As of Nov 3, 2026, CyberBench (v1.1) 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
Automatic (tests or exact checks) — PoC inputs must trigger the OSS-Fuzz crash; patches must fix it under a clean-room grader.

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 · 80.357 · Nov 3, 2026
  2. GPT-6 Sol · 77.976 · Nov 3, 2026
  3. Gemini 4 Argon · 77.857 · Nov 3, 2026
  4. GPT-5.6 Sol · 76.31 · Nov 3, 2026
  5. GPT-6 Luna · 76.25 · Nov 3, 2026
ModelScoreDateKind
Muse Wonder80.3572026-11-03snapshot
GPT-6 Sol77.9762026-11-03snapshot
Gemini 4 Argon77.8572026-11-03snapshot
GPT-5.6 Sol76.312026-11-03snapshot
GPT-6 Luna76.252026-11-03snapshot
Claude Haiku 5.575.4762026-11-03snapshot
Mimo V2.6 Flash75.3572026-11-03snapshot
Kimi K375.2382026-11-03snapshot
DeepSeek V4.1 Flash73.692026-11-03snapshot
GPT-5.6 Luna73.6312026-11-03snapshot
Muse Wonder Full73.6312026-11-03snapshot
DeepSeek V4 Pro 081372.8572026-11-03snapshot
Mimo V2.6 Pro72.8572026-11-03snapshot
Muse Spark 1.372.7382026-11-03snapshot
DeepSeek V4 Flash 073172.0842026-11-03snapshot
GPT-5.6 Terra72.0842026-11-03snapshot
GLM 5.372.0842026-11-03snapshot
Grok 4.571.0122026-11-03snapshot
Claude Fable 5.170.4162026-11-03snapshot
Muse Spark 1.269.5242026-11-03snapshot
Grok 4.769.4642026-11-03snapshot
Muse Spark 1.369.4052026-11-03snapshot
GLM 5.3 Flash67.8572026-11-03snapshot
Qwen3.8 27B67.7982026-11-03snapshot
Grok 4.667.6782026-11-03snapshot
Inkling66.7862026-11-03snapshot
Claude Opus 565.3572026-11-03snapshot
Hy4 Preview65.3572026-11-03snapshot
Minimax M3652026-11-03snapshot
Inkling Small64.1072026-11-03snapshot