HierSVA-B - Formal Core Coverage: leaderboard
HierSVA-B (2026) benchmarks LLM-generated SystemVerilog assertions for hierarchical hardware formal verification on 342 BaseJump STL modules, checked by formal proof with a commercial tool. This board is the coverage, in percent, of the statements and signals in the formal core of each design by the proven assertions; higher is better.
Metric: Formal core coverage (%, C5): union coverage of the design statements and signals in the formal core by the proven assertions on HierSVA-DS (342 hierarchical BaseJump STL modules, depths 0 to 9): static generation of SystemVerilog Assertions from RTL and submodule contracts, checked by formal proof in VC Formal; one run, reasoning enabled, temperature 0, strict JSON output. Source: arxiv.org. Saturation forecast: Around April 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.6 | 53.3 | #178 |
| 2 | GPT-5.5 | 51.4 | #49 |
| 3 | GLM-5.1 | 43.3 | #179 |
| 4 | Qwen 3.6 Max (Preview) (Thinking) | 37.7 | |
| 5 | Claude Opus 4.7 (Thinking) | 34.7 | #79 (Claude Opus 4.7) |
| 6 | Gemini 3.1 Pro (Preview) | 33.8 | #94 |
| 7 | Claude Haiku 4.5 (Thinking) | 33.1 | #568 (Claude Haiku 4.5) |
| 8 | Qwen 3.6 Plus (Thinking) | 29.4 | #222 (Qwen 3.6 Plus) |
| 9 | MiniMax-M2.7 | 26.5 | #518 |
| 10 | GPT-5 Mini | 14.5 | #348 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=hiersva-b-formal-core-coverage · How It Works · Data refreshed daily, snapshot 2026-11-13.