ChartBench

ChartBench evaluates complex visual reasoning over charts across chart types and task combinations.

21rows
all_acc_plusprimary metric
2026-05-06sampled

Metadata

Metrics

Line Acc+, Line Acc, Line COR, Bar Acc+, Bar Acc, Bar COR, Pie Acc+, Pie Acc, Pie COR, Regular Acc+, Regular Acc, Regular COR, Area Acc+, Area Acc, Area COR, Box Acc+, Box Acc, Box COR, Radar Acc+, Radar Acc, Radar COR, Scatter Acc+, Scatter Acc, Scatter COR, Node Acc+, Node Acc, Node COR, Combination Acc+, Combination Acc, Combination COR, Extra Acc+, Extra Acc, Extra COR, All Acc+, All Acc, All COR

Latest Results

Rows ranked by All Acc+, matching the leaderboard's strict aggregate score.

Rank Subject All Acc+ Model Match Provenance Sampled
1 GPT-4O 64.27 GPT-4o
openai-gpt-4o
Imported 2026-05-06
2 GPT-4V 54.39 GPT-4
openai-gpt-4
Imported 2026-05-06
3 Internlm-XComposer-v2 51.34 — Imported 2026-05-06
4 ERNIE 46.95 — Imported 2026-05-06
5 Mini-Gemini 36.54 — Imported 2026-05-06
6 DocOwl-v1.5 31.62 — Imported 2026-05-06
7 Qwen-VL-Chat 28.18 — Imported 2026-05-06
8 mPLUG-Owl-bloomz 26.78 — Imported 2026-05-06
9 LLaVA-v1.5 26.39 — Imported 2026-05-06
10 MiniGPT-v2 23.55 — Imported 2026-05-06
11 ChartLlama 22.26 — Imported 2026-05-06
12 BLIP2 20.24 — Imported 2026-05-06
13 CogAgent 18.07 — Imported 2026-05-06
14 SPHINX 17.89 — Imported 2026-05-06
15 Internlm-XComposer 15.49 — Imported 2026-05-06
16 CogVLM 13.30 — Imported 2026-05-06
17 InstructBLIP 12.49 — Imported 2026-05-06
18 OneChart 12.04 — Imported 2026-05-06
19 Shikra 8.11 — Imported 2026-05-06
20 ChartVLM 6.90 — Imported 2026-05-06
21 VisualGLM 3.79 — Imported 2026-05-06