ArtifactsBench: leaderboard
Benchmark for automated multimodal evaluation of visual and interactive artifact generation from code, using rendered artifacts and checklist-guided MLLM judging.
Metric: AVG (self-reported). Source: benchmarklist.com. Status: saturation imminent. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Ling-3.0-flash | 77 |
| 2 | GPT-5 | 72.55 |
| 3 | Claude Opus 4.1 | 59.76 |
| 4 | Gemini 2.5 Pro | 57.74 |
| 5 | GPT-OSS-120B | 57.69 |
| 6 | Claude Sonnet 4 | 57.28 |
| 7 | Qwen 3 235B A22B (Thinking) | 55.01 |
Interactive version: theaggregate.ai/benchmark?slug=artifactsbench · How It Works · Data refreshed daily, snapshot 2026-09-05.