Issues

• EvalScope integration: one-command evaluation support for CL-bench

  Tencent-Hunyuan/CL-bench#27 · Filter by author Yunnglin opened on Jul 24, 2026

• Official GPT-5.1 judge prompt now rejected by OpenAI pre-inference moderation (400 "flagged as potentially violating usage policy") — official scoring pipeline fails for all tasks

  Tencent-Hunyuan/CL-bench#26 · Filter by author Liu-RK opened on Jul 20, 2026

• Potential instability in official solving rate due to judge sensitivity and all-or-nothing aggregation

  Tencent-Hunyuan/CL-bench#24 · Filter by author wistone opened on Jun 16, 2026

• The Identity Layer: What I'm Building That I Don't See in the Literature

  Tencent-Hunyuan/CL-bench#23 · Filter by author djairjr opened on Jun 2, 2026

• New Work on CL-bench: From Context to Skills: Can Language Models Learn from Context Skillfully?

  Tencent-Hunyuan/CL-bench#20 · Filter by author S1s-Z opened on May 1, 2026
  2

• Request: Update leaderboard with latest model releases (GPT-5.5, Claude Opus 4.7, Gemini 3.1 Pro, etc.)

  Tencent-Hunyuan/CL-bench#19 · Filter by author zgqq opened on Apr 29, 2026
  1

• Data Quality Concerns: Conflicting System Prompts and Poor Context Formatting

  Tencent-Hunyuan/CL-bench#15 · Filter by author wintertee opened on Mar 6, 2026
  1

• GPT-5.2 scores seem off

  Tencent-Hunyuan/CL-bench#13 · Filter by author xeophon opened on Mar 4, 2026
  1

• How to Reproduce Error Type Distribution in Table 3?

  Tencent-Hunyuan/CL-bench#12 · Filter by author Flipped-ciks opened on Feb 27, 2026
  1

• CL-Bench Usage

  Tencent-Hunyuan/CL-bench#8 · Filter by author LLMforScience opened on Feb 9, 2026
  1

• Request: Broader model coverage (Mistral, xAI/Grok, Cerebras, Groq-hosted)

  Tencent-Hunyuan/CL-bench#7 · Filter by author clemlesne opened on Feb 7, 2026
  1

• Improve messages field in CL-bench by adding a delimiter to separate context and task.

  Tencent-Hunyuan/CL-bench#3 · Filter by author 9sxx opened on Feb 4, 2026
  6