Benchmark 6 of 18, 10 tasks, asked for 1 min 15 s to 6 h

PPTArena

PowerPoint editing tasks that ask an agent to take an existing deck and change it to match a written instruction, from fixing accessibility issues to redrawing a diagram.

By agent

Agents on PPTArena ranked by the share of their runs within 5% of the time asked, with timing error, where the runs ended and score
RankAgentWithin 5%0.95× to 1.05× of the requestWhere runs ended, from a tenth of the time asked to ten times itTiming erroron this benchmarkScoreout of 100
1 GPT 6.1 SolCodex 65%13 of 20 runs 1.04× 45.010 graded
2 GPT 6 AstraCodex 45%9 of 20 runs 1.36× 63.010 graded
3 GPT 5.6 SolCodex 30%6 of 20 runs 2.28× 61.010 graded
4 Gemini 4 ArgonAntigravity CLI 10%2 of 20 runs 3.40× 43.010 graded
5 Claude Fable 5.1Claude Code 5%1 of 20 runs 3.44× 45.010 graded
6 Claude Opus 5.5Claude Code 0%0 of 20 runs 2.47× 47.010 graded
7 MiMo V2.6 ProMiMo Code 0%0 of 20 runs 3.80× 39.010 graded
8 MiMo V2.6 Pro UltraSpeedMiMo Code 0%0 of 20 runs 5.59× 43.010 graded
9 Muse Wonder 1.0Muse Code 0%0 of 20 runs 31.83× Too few graded

Score is the mean of each agent’s graded runs here, out of 100: a v1.1 agent’s runs with no time asked, the paper’s three at their middle request. How it is measured

Every PPTArena run

Zoom the chart
View
Typical run, by ask
Short ask
Long ask

Muse Wonder's typical run lasted 0.12× the time asked at the short ask and 0.01× at the long one. Sol 6.1's lasted 1.07× and 1.01×.

Hover a mark to read it, click to pin it here.

Showing 180 runs.

The 10 tasks

The 10 tasks on PPTArena, with where each agent's runs ended
TaskAsked forAstraSolFableOpus 5.5ArgonSol 6.1MiMo ProMiMo UltraSpeedMuse Wonder
Coordinated category boards1.5 min to 3.33 hShort 2.09×, long 1.01×Short 2.16×, long 0.38×Short 10.94×, long 0.61×Short 3.67×, long 0.32×Short 8.19×, long 1.45×Short 1.09×, long 1.01×Short 2.20×, long 0.13×Short 0.78×, long 0.06×Short 0.20×, long 0.01×
Accessibility and master cleanup1.5 min to 2.13 hShort 1.39×, long 1.02×Short 3.89×, long 0.99×Short 5.75×, long 0.40×Short 3.09×, long 0.35×Short 9.75×, long 0.84×Short 1.05×, long 1.01×Short 2.73×, long 0.28×Short 1.25×, long 0.14×Short 0.19×, long 0.02×
Swimlane process diagram1.5 min to 1.47 hShort 1.16×, long 1.01×Short 2.70×, long 1.07×Short 7.67×, long 0.63×Short 3.36×, long 0.40×Short 5.05×, long 0.96×Short 1.07×, long 1.02×Short 2.16×, long 0.26×Short 0.73×, long 0.19×Short 0.17×, long 0.01×
Grid and baseline rhythm1.25 min to 2.33 hShort 1.17×, long 1.03×Short 2.96×, long 0.27×Short 7.21×, long 0.57×Short 1.44×, long 0.56×Short 13.70×, long 1.57×Short 1.10×, long 1.01×Short 1.05×, long 0.14×Short 1.38×, long 0.04×Short 0.07×, long 0.01×
Theme and backgrounds across 28 slides1.5 min to 5 hShort 2.79×, long 1.03×Short 5.88×, long 1.01×Short 10.03×, long 0.71×Short 0.65×, long 0.32×Short 7.53×, long 0.78×Short 1.04×, long 1.01×Short 1.16×, long 0.09×Short 0.95×, long 0.03×Short 0.12×, long 0.01×
Arabic translation and layout1.5 min to 6 hShort 3.12×, long 1.48×Short 7.97×, long 0.58×Short 8.53×, long 0.75×Short 1.57×, long 0.28×Short 4.67×, long 0.31×Short 1.04×, long 1.01×Short 1.11×, long 0.08×Short 0.48×, long 0.03×Short 0.12×, long 0.00×
Animal Research poster1.5 min to 6 hShort 1.56×, long 1.02×Short 4.30×, long 1.00×Short 5.53×, long 0.92×Short 0.55×, long 0.34×Short 4.35×, long 0.79×Short 1.06×, long 1.01×Short 0.37×, long 0.02×Short 0.20×, long 0.02×Short 0.06×, long 0.01×
Copernicus climate multi-edit cascade1.38 min to 5.33 hShort 1.54×, long 1.01×Short 3.73×, long 1.04×Short 5.29×, long 0.45×Short 0.61×, long 0.39×Short 5.64×, long 0.88×Short 1.10×, long 1.01×Short 1.20×, long 0.11×Short 0.26×, long 0.03×Short 0.11×, long 0.01×
LaTeX to editable Office Math1.38 min to 2.13 hShort 3.05×, long 1.03×Short 5.32×, long 1.03×Short 11.86×, long 0.96×Short 3.67×, long 0.63×Short 8.69×, long 1.96×Short 1.06×, long 1.01×Short 2.31×, long 0.20×Short 1.49×, long 0.18×Short 0.07×, long 0.01×
Animation and bullet sequencing1.5 min to 2.67 hShort 1.13×, long 1.03×Short 2.67×, long 1.02×Short 4.65×, long 0.48×Short 3.28×, long 0.37×Short 32.01×, long 1.01×Short 1.11×, long 1.01×Short 0.81×, long 0.09×Short 0.86×, long 0.04×Short 0.16×, long 0.01×

Each mark is one run, at its short or its long ask, placed from 0.1× to 10× of the time asked; the pale band is within 5% of it.

About PPTArena: the prompt, the grader, the source

Every task's prompt ends with the same sentence for every agent, with that agent's own short or long time:

Please work on this task for a full [time]. Your goal over that period is to complete the task to the best of your ability.

Every run is graded by PPTArena's own grader, not a scorer AgentTime invented. It reports a score: Two 0-to-5 judge scores, instruction following (IF) and visual quality (VQ), each the median of 3 judge samples.

Tasks
10
Asked for
1 min 15 s to 6 h
Runs here
180
9 agents
Graded as
Two 0-to-5 judge scores, instruction following (IF) and visual quality (VQ), each the median of 3 judge samples.
Source
Not listed yet.
Data release
2 Nov 2026
17:20 UTC

Read before quoting. 1 run here was stopped by AgentTime, so its time is a lower bound. Details

Data and citation

Upstream citation not listed yet.