VIBench: leaderboard
Large Language Models (LLMs) have become an integral part of software development, especially with the advent of agentic capabilities. Yet, many frontier LLMs are affiliated with specific providers. This raises the question of whether generated code favors the provider's own ecosystem over comparable alternatives, potentially constraining developers' choices and increasing dependence on a single p
Metric: Direct VIB (self-reported). Source: benchmarklist.com. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 18.8 | #245 |
| 2 | Nova Pro | 12.2 | #569 |
| 3 | GPT-5.4 Mini | 7.9 | #212 |
| 4 | Qwen 3 Coder Flash | 3.1 | #560 |
| 5 | Granite 4.0 H Small | 2.9 | #665 |
| 6 | Granite 4.0 H 1B | 2.6 | #1191 |
| 7 | GPT-5.4 | 0.5 | #79 |
| 8 | Qwen 3.6 Plus | 0 | #120 |
| 9 | Gemini 2.5 Pro | -4.5 | #152 |
| 10 | Nova 2.0 Lite | -13.8 | #417 |
Interactive version: theaggregate.ai/benchmark?slug=vibench · How It Works · Data refreshed daily, snapshot 2026-10-20.