Search results

open weight models that don't display when open-weight only is checked

Status: Open.

#341 In pinchbench/skill;

· yieme opened on Apr 15, 2026

Establish target metrics for task difficulty/disparity

Status: Open.

#340 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add security judgment and ethical edge case tasks

Status: Open.

#339 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add task versioning and parameterized templates for contamination resistance

Status: Open.

#338 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Improve LLM-as-judge calibration and reliability

Status: Open.

#337 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add long-context tasks requiring selective retrieval

Status: Open.

#336 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add adversarial/edge-case trap tasks

Status: Open.

#335 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add tasks with intentional ambiguity and underspecification

Status: Open.

#334 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Create multi-step compositional workflow tasks

Status: Open.

#333 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add process/efficiency metrics alongside correctness

Status: Open.

#332 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

Add graduated rubric scoring (partial credit)

Status: Open.

#331 In pinchbench/skill;

· ScuttleBot opened on Apr 15, 2026

new_session: true in sessions frontmatter not implemented — multi-turn tasks cannot isolate conversation context

Status: Open.

#329 In pinchbench/skill;

· HaotianChen616 opened on Apr 15, 2026