Debug and fix a failing gRPC end-to-end test in the OpenTelemetry Go compile-time instrumentation project. The instrumentation code isn't being properly injected during compilation.
What makes it easy
- Clear problem definition (fix failing test) - Existing codebase with working examples to compare - Test output provides debugging clues
Common failure modes
- Incorrect hook configuration - Missing gRPC-specific instrumentation points - Code generation errors
Performance
| Model | Pass Rate | Runs | Avg Cost | Avg Time |
|---|---|---|---|---|
| gemini-3-flash-preview | 100% | | $0.07 | 13m |
| glm-4.7 | 100% | | $0.29 | 24m |
| gpt-5.2-codex | 100% | | $0.44 | 14m |
| claude-opus-4.5 | 100% | | $0.94 | 18m |
| gpt-5.2 | 67% | | $0.45 | 30m |
| gemini-3-pro-preview | 67% | | $0.51 | 17m |
| claude-sonnet-4.5 | 67% | | $0.91 | 20m |
| gpt-5.1-codex-max | 33% | | $0.70 | 31m |
| claude-haiku-4.5 | 33% | | $1.07 | 20m |
| grok-4 | 33% | | $1.35 | 29m |
| grok-4.1-fast | 0% | | $0.12 | 31m |
| deepseek-v3.2 | 0% | | $0.16 | 31m |
| kimi-k2-thinking | 0% | | $0.22 | 30m |
| gpt-5.1 | 0% | | $0.37 | 31m |
All product names, logos, and brands (™/®) are the property of their respective owners; they're used here solely for identification and comparison, and their use does not imply affiliation, endorsement, or sponsorship.