Search results

Inconsistent results for qwen3.5-9b compared with official reported outputs

Status: Open.

#358 In pinchbench/skill;

· Starlight2004 opened on Apr 24, 2026

[BUG] OpenClaw CLI causes task instructions to be overwritten by "Continue where you left off" message

Status: Open.

#345 In pinchbench/skill;

· GruchyC opened on Apr 22, 2026

Cost of Using Claude as a Judge Model per Evaluation

Status: Open.

#342 In pinchbench/skill;

· AmberXu98 opened on Apr 21, 2026

Score variance of different runs

Status: Open.

#103 In pinchbench/skill;

· hanningzhang opened on Apr 4, 2026

reddit research skills

Status: Open.

#97 In pinchbench/skill;

· Ritik-bhaiya opened on Apr 3, 2026

Question: Why does the judge truncate files so aggressively?

Status: Open.

#96 In pinchbench/skill;

· thomas-tharp opened on Apr 2, 2026

[Feature Request] 'Vibe' Benchmark

Status: Open.

#92 In pinchbench/skill;

· joshavant opened on Apr 2, 2026

Why task_11 can test the ClawHub interaction?

Status: Open.

#80 In pinchbench/skill;

· Chaotic-chaos opened on Mar 26, 2026

LLM Judge Prompt Design

Status: Open.

#75 In pinchbench/skill;

· PlanckChang opened on Mar 25, 2026

benchmark: wait for transcript completion instead of assuming subprocess exit means task finished

Status: Open.

#65 In pinchbench/skill;

· jiaxin576 opened on Mar 18, 2026

Multi-domain benchmarking: AgentBench-Live covers 5 domains beyond coding

Status: Open.

#64 In pinchbench/skill;

· jackjin1997 opened on Mar 18, 2026

Leaderboard scoring bias: runs with fewer tasks rank higher than comprehensive runs

Status: Open.

#53 In pinchbench/skill;

· JoeProAI opened on Mar 15, 2026