What do multimodal and grounded LLM benchmarks measure?

They measure whether models can reason over images, charts, documents, and office artifacts instead of only plain-text prompts.

Which benchmarks matter most here?

MMMU-Pro is a strong frontier multimodal reasoning benchmark, while OfficeQA Pro is useful for grounded document and office-style enterprise workflows.

Why is this category separate from knowledge?

A model can know facts in text and still be weak when the information is embedded in a chart, screenshot, PDF, or spreadsheet. This category measures that gap directly.

Multimodal & Grounded

Multimodal & Grounded Benchmarks

Name: Multimodal & Grounded Benchmarks — LLM Leaderboard
Creator: BenchLM.ai

Vision, document, and grounded enterprise workflow benchmarks

MMMU-Pro · OfficeQA Pro

Multimodal and grounded benchmarks test whether a model can read the world as it actually appears in products: screenshots, charts, scanned documents, office files, and mixed visual-text artifacts.

BenchLM.ai currently tracks MMMU-Pro for frontier multimodal reasoning and OfficeQA Pro for grounded enterprise-style document tasks.

This category carries a 12% weight in the overall score. With only two benchmarks, the weight was reduced from 15% until more multimodal evaluations become available, but it remains important for enterprise copilots and document-heavy workflows.

123 models


1 GPT-5.4 Pro OpenAI	Closed	Reasoning	1.05M	91	94%	96%
2 GPT-5.2 Pro OpenAI	Closed	Reasoning	400K	90	96%	96%
3 GPT-5.4 OpenAI	Closed	Reasoning	1.05M	90	95%	96%
4 GPT-5.3 Codex OpenAI	Closed	Reasoning	400K	89	89%	94%
5 GPT-5.2 OpenAI	Closed	Reasoning	400K	88	95%	95%
6 GPT-5.3 Instant OpenAI	Closed	Reasoning	128K	87	95%	95%
7 GPT-5.3-Codex-Spark OpenAI	Closed	Reasoning	256K	87	86%	91%
8 Claude Opus 4.6 Anthropic	Closed	Standard	1M	85	95%	94%
9 GPT-5.2 Instant OpenAI	Closed	Reasoning	128K	85	94%	92%
10 GPT-5.2-Codex OpenAI	Closed	Reasoning	400K	85	84%	92%
11 Gemini 3.1 Pro Google	Closed	Standard	1M	84	95%	95%
12 GPT-5.1-Codex-Max OpenAI	Closed	Reasoning	400K	84	85%	92%
13 Grok 4.1 xAI	Closed	Standard	1M	84	95%	91%
14 Gemini 3 Pro Deep Think Google	Closed	Reasoning	2M	81	95%	95%
15 GPT-5.1 OpenAI	Closed	Reasoning	200K	80	94%	89%
16 GPT-5 (high) OpenAI	Closed	Reasoning	128K	79	93%	85%
17 Claude Sonnet 4.6 Anthropic	Closed	Standard	200K	78	95%	88%
18 GLM-5 (Reasoning) Zhipu AI	Open	Reasoning	200K	78	74%	84%
19 GPT-5 (medium) OpenAI	Closed	Reasoning	128K	78	89%	87%
20 Claude Opus 4.5 Anthropic	Closed	Standard	200K	77	94%	87%
21 Gemini 3 Pro Google	Closed	Standard	2M	77	94%	92%
22 o1-preview OpenAI	Closed	Reasoning	200K	77	72%	80%
23 Claude Sonnet 4.5 Anthropic	Closed	Standard	200K	76	95%	87%
24 Grok 4.1 Fast xAI	Closed	Standard	1M	76	91%	83%
25 Kimi K2.5 (Reasoning) Moonshot AI	Closed	Reasoning	128K	76	72%	77%

Showing 25 of 123

About Multimodal & Grounded Benchmarks

Frontier multimodal reasoning benchmark spanning charts, diagrams, tables, and visual academic problems