Benchmarks.bio - VariantBench — leaderboard
Metric: Pass Rate (%). Source: benchmarks.bio. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 42.1 |
| 2 | GPT-5.6 Sol | 42.1 |
| 3 | Claude Sonnet 5 | 38.7 |
| 4 | GPT-5.6 Terra | 38.7 |
| 5 | Grok 4.5 | 38.1 |
| 6 | Claude Opus 4.7 | 37 |
| 7 | Gemini 3.5 Flash | 35.3 |
| 8 | Gemini 3.1 Pro (Preview) | 34.5 |
| 9 | GPT-5.5 | 30.2 |
| 10 | GPT-5.6 Luna | 29.9 |
| 11 | Grok 4.3 | 20.3 |
Interactive version: aibenchmarks.dev/benchmark?slug=benchmarks-bio-variantbench · How the rankings work · Data refreshed daily, snapshot 2026-07-17.