AI Battle > Rankings > MMMU-Pro: Token Usage

MMMU-Pro: Token Usage Ranking — eval mmmu (2026)

GPT-6 Astra (medium) by OpenAI leads the MMMU-Pro: Token Usage ranking with a score of 1000367. The ranking covers 19 evaluated eval mmmu from a public dataset that AI Battle re-checks daily.

Last updated: . Source: public benchmark datasets and provider pricing, aggregated daily; 19 entries shown.

GPT-6 Astra (medium)1000367GPT-6 Astra (high)1000367GPT-6 Astra (xhigh)1000367GPT-6 Astra (max)1000367Gemini 3.7 Flash (low)2246200GPT-5.6 Sol (max)999093Gemini 3.7 Flash (high)2246200Claude Opus 5 (Adaptive Reasoning, Max Effor1207066Gemini 3.5 Flash-Lite2246200Muse Glimmer (high)1027116Mistral Medium 3.5330058Qwen3.8 27B (xhigh)970778GPT-5.6 Terra (max)1590767Gemini 3.8 Flash (high)2246200Kimi K3 (max)1210646MiniMax-M31359529GPT-5.6 Luna (max)2959753DeepSeek V4.1 Flash (Reasoning, Max Effort)834480Inkling (xhigh)1764121
Every entry in the dataset, scaled from the smallest to the largest published value.

All 19 entries in this dataset

#EntryCreatorValueΔ vs #1
1GPT-6 Astra (medium)OpenAI10003670.0%
2GPT-6 Astra (high)OpenAI10003670.0%
3GPT-6 Astra (xhigh)OpenAI10003670.0%
4GPT-6 Astra (max)OpenAI10003670.0%
5Gemini 3.7 Flash (low)Google2246200124.5%
6GPT-5.6 Sol (max)OpenAI999093-0.1%
7Gemini 3.7 Flash (high)Google2246200124.5%
8Claude Opus 5 (Adaptive Reasoning, Max Effort)Anthropic120706620.7%
9Gemini 3.5 Flash-LiteGoogle2246200124.5%
10Muse Glimmer (high)Meta10271162.7%
11Mistral Medium 3.5Mistral330058-67.0%
12Qwen3.8 27B (xhigh)Alibaba970778-3.0%
13GPT-5.6 Terra (max)OpenAI159076759.0%
14Gemini 3.8 Flash (high)Google2246200124.5%
15Kimi K3 (max)Kimi121064621.0%
16MiniMax-M3MiniMax135952935.9%
17GPT-5.6 Luna (max)OpenAI2959753195.9%
18DeepSeek V4.1 Flash (Reasoning, Max Effort)DeepSeek834480-16.6%
19Inkling (xhigh)Thinking Machines176412176.3%

How the values are distributed

Across the 19 numeric entries in this window the median is 1207066 and the mean is 1433667, so the typical entry sits 206699 above the leader. The window spans 330058 to 2959753.

12821401
8 equal buckets between the smallest (330058) and largest (2959753) value; numbers under bars are entry counts.

10 of 19 entries are at or above the median, and the top three hold 27% of the combined value — read it against the metric below, since a higher number is not "better" for every dataset.

Creators in this ranking

10 creators appear in this window; the three most-represented hold 63% of the entries.

CreatorEntriesHighest valueTop entry
OpenAI72959753GPT-5.6 Luna (max)
Google42246200Gemini 3.7 Flash (low)
Thinking Machines11764121Inkling (xhigh)
MiniMax11359529MiniMax-M3
Kimi11210646Kimi K3 (max)
Anthropic11207066Claude Opus 5 (Adaptive Reasoning, Max Effort)
Meta11027116Muse Glimmer (high)
Alibaba1970778Qwen3.8 27B (xhigh)
DeepSeek1834480DeepSeek V4.1 Flash (Reasoning, Max Effort)
Mistral1330058Mistral Medium 3.5

Where these models appear in other rankings

The same models are measured across multiple independent datasets. Aggregating the 19 entries above against every other AI Battle dataset gives this cross-section, computed by AI Battle and published nowhere upstream:

ModelDatasetsBest placements elsewhere
GPT-6 Astra (medium)12#1/20 in AA-Omniscience Index: Token Usage, #2/20 in Terminal-Bench v4.0: Output Tokens per Task, #6/19 in MMMU-Pro: Score
GPT-6 Astra (high)12#1/20 in AA-Omniscience Index: Score, #2/19 in MMMU-Pro: Score, #3/20 in Terminal-Bench v4.0: Output Tokens per Task
GPT-6 Astra (xhigh)15#1/20 in Terminal-Bench v4.0: Score, #2/20 in Humanity's Last Exam: Output Tokens per Task, #3/20 in AA-Omniscience Index: Score
GPT-6 Astra (max)51#1/19 in MMMU-Pro: Score, #2/20 in AA-Omniscience Index, #2/20 in Artificial Analysis Intelligence Index by Open Weights / Proprietary
Gemini 3.7 Flash (low)3#1/19 in MMMU-Pro: Total Cost to Run, #7/19 in MMMU-Pro: Score
GPT-5.6 Sol (max)56#3/20 in SciCode: Output Tokens per Task, #4/20 in 𝜏³-Banking: Output Tokens per Task, #5/20 in AA-LCR v1.1: Output Tokens per Task
Gemini 3.7 Flash (high)3#5/19 in MMMU-Pro: Total Cost to Run, #5/19 in MMMU-Pro: Score
Claude Opus 5 (Adaptive Reasoning, Max Effort)44#2/20 in AA-Briefcase Elo, #2/20 in GDPval-AA v2 Leaderboard, #2/20 in AA-Briefcase Elo
Gemini 3.5 Flash-Lite44#1/20 in End-to-End Response Time, #1/20 in Output Speed, #1/20 in Time per Intelligence Index Task
Muse Glimmer (high)43#1/20 in Output Tokens per Intelligence Index Task, #1/20 in Cost per Intelligence Index Task, #1/20 in Cost per Intelligence Index Task
Mistral Medium 3.541#2/20 in GDPval-AA v2: Output Tokens per Task, #3/20 in Output Tokens per Intelligence Index Task, #3/20 in Output Tokens per Intelligence Index Task
Qwen3.8 27B (xhigh)43#6/20 in AA-Omniscience Hallucination Rate, #6/20 in 𝜏³-Banking: Score, #8/20 in AA-LCR v1.1: Cost per Task
GPT-5.6 Terra (max)52#7/20 in AA-Briefcase Output Tokens per Task, #8/20 in Humanity's Last Exam: Output Tokens per Task, #9/20 in Output Speed
Gemini 3.8 Flash (high)60#1/11 in Speed, #1/11 in Speed, #2/20 in Output Speed
Kimi K3 (max)60#1/20 in Context Window, #1/20 in AA-LCR v1.1: Score, #1/14 in Model Size: Total and Active Parameters
MiniMax-M357#1/20 in AA-Omniscience Hallucination Rate, #1/20 in 𝜏³-Banking: Output Tokens per Task, #4/20 in Pricing: Cache Hit, Input, and Output
GPT-5.6 Luna (max)58#1/11 in Cost per Task, #1/11 in Cost per Task, #3/20 in Pricing: Cache Hit, Input, and Output
DeepSeek V4.1 Flash (Reasoning, Max Effort)56#3/20 in End-to-End Response Time, #3/20 in Latency: Time To First Answer Token, #2/11 in Cost per Task
Inkling (xhigh)52#3/20 in Context Window, #3/20 in 𝜏³-Banking: Output Tokens per Task, #5/20 in AA-Briefcase Output Tokens per Task

Metric reference

This ranking carries the metrics answerTokens, inputTokens, reasoningTokens from the MMMU-Pro: Token Usage dataset in the tab::eval-mmmu family. Units and the direction of "better" are not declared by the source dataset, so AI Battle does not invent them: values are shown exactly as published, and the ordering is the source dataset's own.

How this ranking is built

Rebuilt on from the public dataset MMMU-Pro: Token Usage (family tab::eval-mmmu). The source lists 19 entries and all 19 are shown here. AI Battle requests the dataset's rows, keeps the ordering the source publishes, and computes only the statistics on this page: median, mean, spread, the #1–#2 gap, the creator breakdown and the raw-data file. Nothing on this page is generated text about the entries — the numbers come from the source.

Refresh cadence: every deploy re-pulls the dataset, so this page tracks the source's own update cycle. Corrections: see the methodology page.

Limitations

Raw data

Machine-readable copy of this table, with the statistics below and the source provenance: /data/rankings/eval-mmmu-mmmu-pro-token-usage.json. Free to reuse with attribution to AI Battle and the source dataset.

FAQ — MMMU-Pro: Token Usage

What is the best model for MMMU-Pro: Token Usage in 2026?

GPT-6 Astra (medium) by OpenAI currently leads the MMMU-Pro: Token Usage ranking with a score of 1000367, the highest value in this dataset across eval mmmu. AI Battle aggregates the dataset from its public source and re-checks it daily.

How often is the MMMU-Pro: Token Usage ranking updated?

The MMMU-Pro: Token Usage ranking is rebuilt daily from the upstream public dataset. Scores are the source dataset's own published values, so positions track whatever that source last released — each ranking page names the dataset it comes from.

Which models rank in the top 3 for MMMU-Pro: Token Usage?

The top 3 are: 1. GPT-6 Astra (medium) (1000367), 2. GPT-6 Astra (high) (1000367), 3. GPT-6 Astra (xhigh) (1000367).

All AI Battle Rankings

Directories