Knowledge work
AA-Briefcase
Realistic multi-file knowledge-work projects (slides, spreadsheets, memos) scored on rubric, analytical quality, and presentation.
As of 13 August 2026. Conditional pack. Artificial Analysis Briefcase, observed Jul 2026. Rubric, analytical-quality Elo, and presentation Elo stay separate.
AA-Briefcase scores realistic knowledge-work files: slides, spreadsheets, memos. Rubric pass, analytical-quality Elo, and presentation Elo are different estimands. A combined Briefcase Elo is not used as a winner here.
This is a third-party harness. It is not an Antidote engagement and not a client memo sample.
- Scale
- Artificial Analysis Briefcase harness, observed Jul 2026
- Provenance
- Artificial Analysis, GPT-5.6 article 9 Jul 2026 and Fable 5 model page. Third-party harness, not an Antidote evaluation.
- Contamination
- Unknown. Private expert-built projects in the AA harness. Not a public CSV.
- Cannot establish
- A client memo score, a vendor product score, or an Antidote run.
Metrics, kept separate
- Rubric score
- Analytical-quality Elo
- Presentation Elo
Third-party leaderboard · not an Antidote run · not a product score
Artificial Analysis GPT-5.6 article, 9 Jul 2026. Rubric score stays separate from analytical-quality Elo (Fable 5 1764, Sol 1592). Presentation Elo is described as Sol leading. Third-party harness, not an Antidote run.
| System | Rubric score | Analytical-quality Elo |
|---|---|---|
| Claude Fable 5 max | 56% | 1764 |
| GPT-5.6 Sol max | 42% | 1592 |
What the figures bound
Claude Fable 5 (max) rubric score is 56%. GPT-5.6 Sol (max) is 42%. Analytical-quality Elo is 1764 for Fable 5 and 1592 for Sol. AA states Sol has the highest Presentation Elo of any model they recorded, while Fable 5 still leads the overall Briefcase comparison on rubric and analytical quality.
Presentation Elo is not memo correctness. Rubric score is not an independent product test.
What this page does not conclude
- A client memo, slide, or spreadsheet product score
- An Antidote Briefcase run
- A single Briefcase winner from mixed Elo and rubric
- Adoption or seat-compression value