Skip to main content
BenchLM

GDP.pdf mean criteria pass rate without tools (GDP.pdf (no tools))

We show this table for reference; we do not rank on it.

Data verified 34 confirmed releases in the last 30 daysFollow model changes

Professional document understanding over 100 real-world PDFs from ten domains.

Benchmark score on GDP.pdf (no tools) — September 27, 2026

We compile the GDP.pdf (no tools) rows from provider self-reports. Claude Opus 5 leads the table at 83.4%, followed by dots3-note Preview (60.7%). We do not use these results to rank models overall.

2 modelsMultimodal & GroundedCurrentDisplay onlyUpdated September 27, 2026

Benchmark score table (2 models)

Score
1
Claude Opus 5Anthropic · Closed
83.4%
2
dots3-note PreviewDots Studio · Open weight
60.7%

About GDP.pdf (no tools)

Year

2026

Tasks

100 professional document prompts

Format

Mean criteria pass rate

Difficulty

Professional document reasoning

Section 8.12.4 reports a five-run mean criteria pass rate from Anthropic's corrected internal harness without tools, using Opus 4.7 as judge.

Freshness and provenance

Version

GDP.pdf (no tools) 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does GDP.pdf (no tools) measure?

Professional document understanding over 100 real-world PDFs from ten domains.

Which model scores highest on GDP.pdf (no tools)?

Claude Opus 5 by Anthropic currently leads with a score of 83.4% on GDP.pdf (no tools).

How many models are evaluated on GDP.pdf (no tools)?

2 AI models have been evaluated on GDP.pdf (no tools) on BenchLM.

Compare top models on GDP.pdf (no tools)

Last updated: September 27, 2026 · BenchLM version GDP.pdf (no tools) 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.