Benchmark Almanac

InfographicVQA

Kimi K2.5 92.6% leads since Jan 2026 In use

Models answer about 30,000 questions on about 5,000 infographics that mix text, layout, charts, and graphics. Scoring is ANLS.

In the authors’ words

Infographics are documents designed to effectively communicate information using a combination of textual, graphical and visual elements. In this work, we explore the automatic understanding of infographic images by using Visual Question Answering technique.To this end, we present InfographicVQA, a new dataset that comprises a diverse collection of infographics along with natural language questions and answers annotations. The collected questions require methods to jointly reason over the document layout, textual content, graphical elements, and data visualizations. We curate the dataset with emphasis on questions that require elementary reasoning and basic arithmetic skills. Finally, we evaluate two strong baselines based on state of the art multi-modal VQA models, and establish baseline performance for the new task. The dataset, code and leaderboard will be made available at http://docvqa.org

Abstract of InfographicVQA (arXiv:2104.12756, CC0)
Made by Computer Vision Center (UAB) and IIIT Hyderabad Released Apr 2021 Documents, charts and text in images All tasks public Last in a launch post Aug 2026 Last model added Aug 2026
Level
School Students, from grade school to high school
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Short answer A number, word or short string matched to a key.
Paper docvqa.org
22
Models scored in lab cards
15
Launch posts list it, from 9 labs
14
System cards report it, from 8 labs

Kimi K2.5 leads with 92.6% among models released by Aug 2026

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

The newest model with a score here came out in Aug 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

22 models. Scroll sideways for the rest.

  1. 92.6 Kimi K2.5
  2. 92.0 Gemma 4 31B IT
  3. 91.2 Seed1.5-VL
  4. 89.5 Qwen3-VL-235B-A22B
  5. 89.3 Gemma 4 26B A4B
  6. 88.0 Seed1.6-Thinking
  7. 87.3 Qwen2.5-VL-72B
  8. 86.8 Nemotron 3 Nano Omni 30B-A3B
  9. 82.9 Command A Vision
  10. 82.0 InternVL3.5-241B-A28B
  11. 81.0 Gemini 1.5 Pro
  12. 80.4 Nemotron Nano 12B v2 VL
  13. 80.3 Gemini 1.0 Ultra
  14. 77.4 Llama-3.1-Nemotron-Nano-VL-8B-V1
  15. 72.7 Gemini 1.5 Pro
  16. 72.7 Phi-4-multimodal
  17. 70.6 Gemma 3 27B
  18. 70.2 LFM2.5-VL-3B
  19. 65.2 North Micro Vision Instruct
  20. 64.9 Gemma 3 12B
  21. 58.7 LFM2-VL-1.6B
  22. 50.0 Gemma 3 4B
  1. 1 Kimi K2.5 92.6%
  2. 2 Gemma 4 31B IT 92%
  3. 3 Seed1.5-VL 91.2%
  4. 4 Qwen3-VL-235B-A22B 89.5%
  5. 5 Gemma 4 26B A4B 89.3%
  6. 6 Seed1.6-Thinking 88%
  7. 7 Qwen2.5-VL-72B 87.3%
  8. 8 Nemotron 3 Nano Omni 30B-A3B 86.8%
  9. 9 Command A Vision 82.9%
  10. 10 InternVL3.5-241B-A28B 82%
  11. 11 Gemini 1.5 Pro May 2024 81%
  12. 12 Nemotron Nano 12B v2 VL 80.4%
  13. 13 Gemini 1.0 Ultra 80.3%
  14. 14 Llama-3.1-Nemotron-Nano-VL-8B-V1 77.4%
  15. 15 Gemini 1.5 Pro Feb 2024 72.7%
  16. 16 Phi-4-multimodal 72.7%
  17. 17 Gemma 3 27B 70.6%
  18. 18 LFM2.5-VL-3B 70.2%
  19. 19 North Micro Vision Instruct 65.2%
  20. 20 Gemma 3 12B 64.9%
  21. 21 LFM2-VL-1.6B 58.7%
  22. 22 Gemma 3 4B 50%
#ModelScoreSettingDateSource
1 Kimi K2.5 92.6% Jan 2026 Model card
2 Gemma 4 31B IT 92% Jul 2026 Model card
3 Seed1.5-VL 91.2% May 2025 Model card
4 Qwen3-VL-235B-A22B 89.5% Sep 2025 Model card
5 Gemma 4 26B A4B 89.3% Jul 2026 Model card
6 Seed1.6-Thinking 88% Jun 2025 Model card
7 Qwen2.5-VL-72B 87.3% Feb 2025 Model card
8 Nemotron 3 Nano Omni 30B-A3B 86.8% Apr 2026 Model card
9 Command A Vision 82.9% Jul 2025 Model card
10 InternVL3.5-241B-A28B 82% Aug 2025 Model card
11 Gemini 1.5 Pro May 2024 81% May 2024 Model card
12 Nemotron Nano 12B v2 VL 80.4% Nov 2025 Model card
13 Gemini 1.0 Ultra 80.3% Dec 2023 Model card
14 Llama-3.1-Nemotron-Nano-VL-8B-V1 77.4% Jun 2025 Model card
15 Gemini 1.5 Pro Feb 2024 72.7% Feb 2024 Model card
16 Phi-4-multimodal 72.7% Feb 2025 Model card
17 Gemma 3 27B 70.6% Mar 2025 Model card
18 LFM2.5-VL-3B 70.2% Aug 2026 Model card
19 North Micro Vision Instruct 65.2% Aug 2026 Model card
20 Gemma 3 12B 64.9% Mar 2025 Model card
21 LFM2-VL-1.6B 58.7% Aug 2025 Model card
22 Gemma 3 4B 50% Mar 2025 Model card

5 models have held number one since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind Qwen ByteDance Seed Moonshot AI
Lab claims Passed 85% in Feb 2025, by Qwen2.5-VL-72B
Kimi K2.5 92.6%

The best score passed 85% in Feb 2025

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Kimi K2.5 Seed1.5-VL Gemini 1.5 Pro 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Kimi K2.5 Gemini 1.5 Pro

2 labs put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

OpenAI
Google DeepMind
Microsoft
ByteDance Seed
NVIDIA
Cohere
Liquid AI
Qwen
Moonshot AI
DateLaunch postScoreAs the post writes it
12 Aug 2026 LFM2.5-VL-3B 70.2 InfographicVQA (val)
12 Aug 2026 North Micro Vision Instruct 65.2 InfoVQA VAL
27 Jan 2026 Kimi K2.5 92.6 InfoVQA (val)
23 Sep 2025 Qwen3-VL-235B-A22B 89.5 InfoVQA_TEST
12 Aug 2025 LFM2-VL-1.6B 58.7 InfoVQA (Val)
31 Jul 2025 Command A Vision 82.9 InfoVQA
25 Jun 2025 Seed1.6-Thinking 88 Infographic VQA
3 Jun 2025 Llama-3.1-Nemotron-Nano-VL-8B-V1 77.4 InfoVQA Val
12 May 2025 Seed1.5-VL 91.2 InfographicVQA
12 Mar 2025 Gemma 3 27B 70.6 InfoVQA (IT); PT InfoVQA (val) = 59.4
26 Feb 2025 Phi-4-multimodal 72.7 InfoVQA
14 May 2024 Gemini 1.5 Pro 81 InfographicVQA (test), 0-shot (ANLS)
15 Feb 2024 Gemini 1.5 Pro 72.7 InfographicVQA (test), 0-shot (ANLS)
6 Dec 2023 Gemini 1.0 Ultra 80.3 InfographicVQA (test)
14 Mar 2023 GPT-4 75.1 Infographic VQA ANLS score (test) 0-shot (pixel-only)

14 system cards report it, from 8 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
24 Jul 2026 Gemma 4 31B IT technical report 92 InfographicVQA, thinking, max resolution (Table 6) Capabilities
24 Jul 2026 Gemma 4 26B A4B technical report 89.3 InfographicVQA, thinking, max resolution (Table 6) Capabilities
27 Apr 2026 Nemotron 3 Nano Omni 30B-A3B technical report 86.8 InfoVQA (Test), reasoning on, 83.6 reasoning off (Table 7) Capabilities
30 Jan 2026 Kimi K2.5 technical report 92.6 InfoVQA (test) (Table 4) Capabilities
28 Nov 2025 LFM2-VL-1.6B technical report 58.4 InfoVQA (val), LFM2-VL-1.6B (Table 8) Capabilities
26 Nov 2025 Qwen3-VL-235B-A22B technical report 89.5 InfoVQA test (Table 2) Capabilities
6 Nov 2025 Nemotron Nano 12B v2 VL technical report 80.4 InfoVQA (Test), reasoning on (Table 3). Reasoning off 79.4 Capabilities
25 Aug 2025 InternVL3.5-241B-A28B technical report 82 InfoVQA (test) (Table 4) Capabilities
11 May 2025 Seed1.5-VL technical report 91.2 InfographicVQA, thinking (Table 6). Non-thinking 89.3 Capabilities
25 Mar 2025 Gemma 3 4B technical report 50 InfoVQA, IT, P&S applied (Table 16) Capabilities
25 Mar 2025 Gemma 3 27B technical report 70.6 InfoVQA, IT, P&S applied (Table 16) Capabilities
25 Mar 2025 Gemma 3 12B technical report 64.9 InfoVQA, IT, P&S applied (Table 16) Capabilities
3 Mar 2025 Phi-4-multimodal technical report 72.7 InfoVQA (test) (Table 1). Speech-prompt ShareGPT4o InfoVQA 63.7 (Table 2) Capabilities
19 Feb 2025 Qwen2.5-VL-72B technical report 87.3 InfoVQA test (Table 5) Capabilities

More in Vision

All 234 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. GPT-5.1 85.4%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. Kimi K3.1 Pro 84.1%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance. Command A Vision 90.9%
  • DocVQA In use Models answer 50,000 questions about 12,000 scanned document images by reading their text and layout. Scoring is ANLS, a string similarity to reference answers. Seed1.5-VL 96.9%