Inspect AI
by UK AI Safety Institute · open-source · Last verified 2026-03-17
Open-source framework for evaluating LLMs developed by the UK AI Safety Institute. Provides standardized evaluation tools for assessing model capabilities, safety, and alignment with reproducible benchmarks.
https://inspect.ai-safety-institute.org.uk ↗C
C—Below Average
Adoption: CQuality: AFreshness: A+Citations: CEngagement: F
Specifications
- License
- MIT
- Pricing
- open-source
- Capabilities
- model-evaluation, safety-assessment, benchmark-reproduction, custom-evaluations, scoring
- Integrations
- openai, anthropic, hugging-face
- Use Cases
- safety-evaluation, model-benchmarking, capability-assessment, alignment-testing
- API Available
- Yes
- SDK Languages
- python
- Deployment
- self-hosted
- Rate Limits
- N/A (open-source)
- Data Privacy
- Self-hosted, user-managed
- Tags
- evaluation, safety, benchmarking, uk-aisi
- Added
- 2026-03-17
- Completeness
- 100%
Index Score
44.25Adoption
40
Quality
85
Freshness
90
Citations
45
Engagement
0