BenchmarksGDP.pdf w/ python
GDP.pdf w/ python
Completing PDF-based knowledge-work tasks with Python, measured by the mean share of rubric criteria passed per task.
Test definition
- Author / evaluator
- Reported by Moonshot AI
- Version / family
- GDP.pdf w/ python / GDP.pdf w/ python
- Evidence available
- 5 results in 1 comparison groups
How the test is scored
Python receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf.
Comparison notes
Publisher-reported results, read 24 October 2026. Compare reported values only within the same source row; tool settings, subsets, versions, judges and metrics must not be combined.
A different task split, tool policy, judge or reasoning effort can change the result. Choose one comparison group below.
Results by evaluator
Select a publication and metric to compare the recorded field.
selfKimi K3.1 Pro: complete release evaluation · GDP.pdf w/ python
- 85.7%
- 85.5%
- 83.9%
- 83.8%
- 73.0%
Bars start at zero. 5 reported results in this comparison group; ties share a rank. Source ↗
All observations
Every numeric result, with its source and reported run conditions.
5 results
| Model | Result (%) | Source | Run |
|---|---|---|---|
| Kimi K3.1 Pro | 85.7 | self | DetailsPython receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf. Read 24 October 2026. All 47 source rows and six comparison columns are retained, with text-only HLE and both ZeroBench metrics split explicitly. Kimi uses max reasoning effort, temperature 1.0 and top-p 0.95 unless a row states otherwise; proprietary comparison columns are labelled max with fallback. Comparison figures include quoted external results and mixed harnesses. Missing cells remain null and do not enter ranks. Lower hallucination rates rank first; every other rank compares larger values within the same row. Terminal-Bench Science’s September DeepSeek score is attributed separately according to footnote 2. These publication claims do not populate independent boards. Highspeed has no independently identified results and inherits none of these scores.
|
| Claude Fable 5.1 | 83.8 | self | DetailsPython receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf. Read 24 October 2026. All 47 source rows and six comparison columns are retained, with text-only HLE and both ZeroBench metrics split explicitly. Kimi uses max reasoning effort, temperature 1.0 and top-p 0.95 unless a row states otherwise; proprietary comparison columns are labelled max with fallback. Comparison figures include quoted external results and mixed harnesses. Missing cells remain null and do not enter ranks. Lower hallucination rates rank first; every other rank compares larger values within the same row. Terminal-Bench Science’s September DeepSeek score is attributed separately according to footnote 2. These publication claims do not populate independent boards. Highspeed has no independently identified results and inherits none of these scores.
|
| GPT-6 Astra | 85.5 | self | DetailsPython receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf. Read 24 October 2026. All 47 source rows and six comparison columns are retained, with text-only HLE and both ZeroBench metrics split explicitly. Kimi uses max reasoning effort, temperature 1.0 and top-p 0.95 unless a row states otherwise; proprietary comparison columns are labelled max with fallback. Comparison figures include quoted external results and mixed harnesses. Missing cells remain null and do not enter ranks. Lower hallucination rates rank first; every other rank compares larger values within the same row. Terminal-Bench Science’s September DeepSeek score is attributed separately according to footnote 2. These publication claims do not populate independent boards. Highspeed has no independently identified results and inherits none of these scores.
|
| Claude Opus 5.5 | 83.9 | self | DetailsPython receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf. Read 24 October 2026. All 47 source rows and six comparison columns are retained, with text-only HLE and both ZeroBench metrics split explicitly. Kimi uses max reasoning effort, temperature 1.0 and top-p 0.95 unless a row states otherwise; proprietary comparison columns are labelled max with fallback. Comparison figures include quoted external results and mixed harnesses. Missing cells remain null and do not enter ranks. Lower hallucination rates rank first; every other rank compares larger values within the same row. Terminal-Bench Science’s September DeepSeek score is attributed separately according to footnote 2. These publication claims do not populate independent boards. Highspeed has no independently identified results and inherits none of these scores.
|
| DeepSeek V4.1 Flash 1017 | 73.0 | self | DetailsPython receives the original PDF in an IPython sandbox. Five-run task-macro rubric pass rate, judged by Gemini-3.5-Flash medium; this is not strict all-pass GDP.pdf. Read 24 October 2026. All 47 source rows and six comparison columns are retained, with text-only HLE and both ZeroBench metrics split explicitly. Kimi uses max reasoning effort, temperature 1.0 and top-p 0.95 unless a row states otherwise; proprietary comparison columns are labelled max with fallback. Comparison figures include quoted external results and mixed harnesses. Missing cells remain null and do not enter ranks. Lower hallucination rates rank first; every other rank compares larger values within the same row. Terminal-Bench Science’s September DeepSeek score is attributed separately according to footnote 2. These publication claims do not populate independent boards. Highspeed has no independently identified results and inherits none of these scores.
|
Reproduction
Public execution instructions have not been verified in this catalogue. This status does not claim that the benchmark is closed or impossible to run.
Execution documentation ↗