twinkle-eval
High-performance LLM evaluation framework with parallel API calls — up to 17× faster than sequential tools. Supports box, math, and logit-based evaluation.
View on Github
Activity
- Latest release
- 8h ago
- Total releases
- 8
- Cadence
- ~2 days
- Last 12 months
- 8
Reach
- Stars
- 110
Details
- License
- MIT
- First release
- Mar 31, 2026
Releases
| Version | Released | |
|---|---|---|
2.10.0
minor
| ||
2.9.0
minor
| ||
2.8.1
patch
| ||
2.8.0
minor
| ||
2.7.1
patch
| ||
2.7.0
minor
| ||
2.6.0
minor
| ||
2.5.0
initial
|