Benchmark Almanac

PosterBench

Not in model cards

Agents turn research papers into posters. Programmatic checks and visual model judgments score seven content and design dimensions, with ceilings for serious rendering or layout failures.

The benchmark has a 100-paper track and a fixed 10-paper subset. Its star count belongs to the broader AutoDesign project.

Released
August 2026
Kind
Documents and spreadsheets
Level
University
Scored
Graded by a model
Answers
Agent task
Tasks
All public
Measures
Capability
Subject
Arts
Language
English
In and out
Text, Documents → Documents, Actions
Interaction
Agentic
Environment
Terminal
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Agents

In the authors’ words

Transforming multimodal sources into condensed and structured media outputs can be fundamentally conceptualized as a long-horizon agentic process centered on a model-harness system. While an ideal harness system should align with human design priors and accumulate reusable experience through empirical exploration to drive recursive self-improvement, existing paradigms remain static and fall short of this capability. In this paper, we present AutoDesign, a framework that aligns with human design priors, where a meta-harness optimizer guides a code agent to recursively improve harness based on rollout feedback. To instantiate and evaluate this framework, we focus on the academic paper-to-poster generation task and introduce PosterBench, comprising a 100-paper Main Track spanning five disciplines and PosterBench-mini, a shared 10-paper subset for controlled evaluation. On the PosterBench Main Track, AutoDesign achieves the highest score of 78.32, surpassing the closed-source commercial system Claude Design by 7.45 points. Across seven controlled code-agent-model configurations, integrating the learned DesignHarness consistently improves performance, increasing the average PosterBench Score from 54.99 to 67.39 (+12.4%). In a fully autonomous long-horizon loop, it executes 253 tool calls and 11 editing turns within 40 minutes for under $3, reaching average conference-poster quality in human evaluation. A system-blind human study further demonstrates that AutoDesign achieves the highest human preference among evaluated systems.

Abstract of AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design (arXiv:2608.13560, CC0)
14
Entries on its own leaderboard, 0 models we track

AutoDesign / DesignHarness / Claude Code + Claude Opus 4.8 tops the leaderboard with 78.3 No scores yet

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. Agents built around a model, such as AutoDesign / DesignHarness / Claude Code + Claude Opus 4.8, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

Source: Official PosterBench leaderboard

#EntryScoreSettingDateSource
1 AutoDesign / DesignHarness / Claude Code + Claude Opus 4.8 78.3 100-paper main track May 2026 Official leaderboard
2 AutoDesign / DesignHarness / Codex + GPT-5.5 78.0 100-paper main track Apr 2026 Official leaderboard
3 Codex / Codex + GPT-5.5 73.4 100-paper main track Apr 2026 Official leaderboard
4 Claude Design / Claude Design / Claude Code + Claude Opus 4.8 70.9 100-paper main track May 2026 Official leaderboard
5 Claude Code / Claude Code + Claude Opus 4.8 70.0 100-paper main track May 2026 Official leaderboard
6 OpenDesign / OpenDesign / Claude Code + Claude Opus 4.8 69.5 100-paper main track May 2026 Official leaderboard
7 OpenDesign / OpenDesign / Codex + GPT-5.5 62.2 100-paper main track Apr 2026 Official leaderboard
8 Seed 2.1 61.1 100-paper main track, Doubao / Claude Code Official leaderboard
9 PosterGen + Claude Opus 4.8 56.7 100-paper main track May 2026 Official leaderboard
10 GLM / Claude Code + GLM-5.2 52.2 100-paper main track Jun 2026 Official leaderboard
11 Kimi / Claude Code + Kimi K2.7 Code 51.5 100-paper main track Jun 2026 Official leaderboard
12 Any2Poster + Claude Opus 4.8 49.1 100-paper main track May 2026 Official leaderboard
13 DeepSeek / Claude Code + DeepSeek-V4-Pro 46.0 100-paper main track Apr 2026 Official leaderboard
14 Paper2Poster + Claude Opus 4.8 44.6 100-paper main track May 2026 Official leaderboard
#ModelScoreSettingDateSource

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Economic work

All 210 in Economic work
  • JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
  • GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate. Seed2.1 Pro 87.9%
  • GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
  • AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
  • Agents' Last Exam v1.1 In use Agents complete 122 public long-horizon professional workflow tasks across 55 sub-industries. Deterministic graders check outcomes against hidden references. GPT-6.2 Astra 47.5%
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Claude Fable 5 75.9%

Data audit

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (14 entries).
  • Pass Models scored 14 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on raw.githubusercontent.com: 13 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check