activeVision

GDP.pdf

Metric: Accuracy (higher is better)Introduced: 2026

Visual document comprehension benchmark using economic and financial PDF charts and figures without tools.

Models ranked

7

tracked on this benchmark

Score band

81.6 – 24.9

best → lowest tracked

Snapshot trend

+0.60

Aug 13 → Oct 23 · 1 models

Leaderboard

Tracked models ranked by Accuracy (higher is better).

Compare candidates
#Model variant and provenanceScore
1
Claude Sonnet 5
Version: GDP.pdf with Python tools and image cropping tool; internal harness, Opus 4.7 judge, mean criteria pass rate averaged over 5 runsHarness: Not recordedEvaluator: Not recordedObserved: Jun 30, 2026Confidence: Not recordedSource
81.6
2
Kimi K3.1 Pro
Version: GDP.pdf vision, no tools; strict pass rate; five-run mean; max reasoning effortHarness: Not recordedEvaluator: Not recordedObserved: Oct 23, 2026Confidence: Not recordedSource

Notes: Moonshot self-reported launch table, row 'GDP.pdf', column 'Kimi K3.1 Pro (max)'. Settings: reasoning effort max, temperature 1.0, top-p 0.95 (blog footnote). Moonshot-run (Gemini-3.5-Flash judge).

34.6
3
Gemini 3.7 Flash
Version: GDP.pdf; Gemini self-computedHarness: Not recordedEvaluator: Not recordedObserved: Aug 13, 2026Confidence: Not recordedSource
34.0
4
Muse Wonder Full

Configuration: Muse Wonder Full (max)

Version: Not recordedHarness: Not recordedEvaluator: MetaObserved: Oct 12, 2026Confidence: self-reportedSource

Notes: Meta first-party benchmark table on the Muse Wonder model page; methodology in the Muse Wonder model card.

33.2
5
Muse Wonder

Configuration: Muse Wonder (max)

Version: Not recordedHarness: Not recordedEvaluator: MetaObserved: Oct 12, 2026Confidence: self-reportedSource

Notes: Meta first-party benchmark table on the Muse Wonder model page; methodology in the Muse Wonder model card.

30.8
6
Claude Fable 5
Version: GDP.pdf vision, no toolsHarness: Not recordedEvaluator: Not recordedObserved: Jun 9, 2026Confidence: Not recordedSource

Notes: Unstarred in Anthropic's shared Fable 5/Mythos 5 launch table, so this is a Claude Fable 5 score.

29.8
7
GPT-5.5
Version: GDP.pdf vision, no toolsHarness: Not recordedEvaluator: Not recordedObserved: Jun 9, 2026Confidence: Not recordedSource

Notes: GPT-5.5 score from Anthropic's Fable 5 comparison table.

24.9

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

Related benchmarks

ARC-AGI-2 High
ARC-AGI-2 — High Effort
Blueprint-Bench 2
HR-Bench 4K
Vision Arena
LMArena Vision Arena
SimpleVQA with Search Tool
SWE-bench Multimodal

Last reviewed: Jun 15, 2026

Resources

Website