CG-AL-X248

hard Codeunits & Business Logic content b80146bedf0e…

Description

Per-model results

Models that have attempted this task
ModelAttempt 1Attempt 2Avg scoreTokens / attemptRuns
GPT-6.1 Sol✓—100.0 / 1004.3k / 1.8k3
Gemini 3.8 Flash (OpenRouter)✓—100.0 / 1004.9k / 30k3
GPT-5.6 Sol✓✓85.0 / 1004.8k / 2.4k3
GPT-6 Sol✓✓85.0 / 1004.8k / 1.6k3
Claude Opus 5✗✓81.3 / 1008.3k / 6.7k3
Claude Sonnet 5✗✓81.3 / 1008.7k / 8.7k3
Claude Opus 5.5✗✓75.0 / 1009.9k / 3.1k3
GPT-5.6 Luna✗✓75.0 / 1005.1k / 2.5k3
GPT-5.6 Terra✗✓75.0 / 1005.6k / 2.0k3
MiniMax M3✗✓68.8 / 1007.2k / 17k3
Claude Haiku 4 5✗✗62.5 / 1009.5k / 2.0k3
GPT-6 Luna✗✓58.3 / 1005.5k / 2.9k3
GLM 5.3 (OpenRouter)✗✓40.6 / 1004.2k / 57k2