DV-World - DV-Sheet Create: leaderboard
Native spreadsheet chart creation on DV-World (2026), run in the DV-World-Agent framework: an MLLM judge rubric score for reliability, appropriateness and aesthetics combined with coverage of the gold data values. Percent; higher is better.
Metric: Chart creation overall (%): weighted combination of the MLLM rubric score (reliability, appropriateness, aesthetics) and table coverage of gold values for native spreadsheet chart creation tasks, run in the DV-World-Agent framework (data manipulation, multimodal perception and proactive interaction tools); expert rubrics applied by a Gemini-2.5-Flash judge where rubric-scored, means over evaluation trials; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 36.07 | #123 |
| 2 | GPT-5.2 | 34.43 | #190 |
| 3 | GPT-4.1 | 29.98 | #506 |
| 4 | DeepSeek V3.2 | 28.31 | #425 |
| 5 | Kimi K2 (Thinking) | 26.94 | |
| 6 | Gemini 2.5 Pro | 25.75 | #280 |
| 7 | Grok 4 | 21.29 | #321 |
| 8 | GLM-4.7 | 19.16 | #383 |
| 9 | Qwen 3 Coder Plus | 17.04 | #729 |
| 10 | Qwen 3 235B A22B | 14.32 | #633 |
| 11 | Qwen 3 8B | 1.52 | #1890 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=dv-world-dv-sheet-create · How It Works · Data refreshed daily, snapshot 2026-11-22.