Agent Experience Score for Auth0 | Auth0
Can your AI agent implement Auth0?
See how Auth0 works with your favorite LLM and framework
Auth0 Agent Experience Score
Test how Auth0 works with different models, frameworks, and configurations.
| Feature | Quickstarts |
|---|---|
| Framework | All frameworks |
| Model | Framework |
| Agentic score | Agentic + Auth0 tools |
| Grade | Est. Cost |
| Est. Time |
Grader Results Summary
Claude Haiku 5.5
Framework: Express API
Score: 80.5%
Tool Calls: 14
Cost: $0.74
Time: 1m 6s
Grading Breakdown:- Setup Friction: 10.8/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 11.1/12 pts
- Error Recovery: 5.6/7 pts
- Docs Quality: 7/7 pts
- Correctness: 23.1/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 95/100 pts
GPT-5.6 Luna
Framework: React
Score: 97.8%
Tool Calls: 9
Cost: $0.13
Time: 43s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
Claude Haiku 5.5
Framework: FastAPI
Score: 62.4%
Tool Calls: 13
Cost: $0.52
Time: 54s
Grading Breakdown:- Setup Friction: 10.8/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 11.1/12 pts
- Error Recovery: 5.6/7 pts
- Docs Quality: 7/7 pts
- Correctness: 17.9/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 89/100 pts
GPT-5.6 Luna
Framework: Angular
Score: 95.4%
Tool Calls: 12
Cost: $0.15
Time: 48s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
Claude Haiku 5.5
Framework: Express
Score: 95.1%
Tool Calls: 13
Cost: $1.06
Time: 1m 38s
Grading Breakdown:- Setup Friction: 10.8/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 11.1/12 pts
- Error Recovery: 5.6/7 pts
- Docs Quality: 7/7 pts
- Correctness: 22.1/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 94/100 pts
GPT-5.6 Luna
Framework: React Native
Score: 94.3%
Tool Calls: 18
Cost: $0.27
Time: 1m 16s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
Gemini 3.1 Pro
Framework: React
Score: 96.7%
Tool Calls: 8
Cost: $0.28
Time: 50s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
Gemini 3.1 Pro
Framework: SPA JS
Score: 100%
Tool Calls: 8
Cost: $0.28
Time: 41s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
GPT-5.6 Luna
Framework: Expo
Score: 89%
Tool Calls: 14
Cost: $0.32
Time: 2m 24s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
GPT-5.6 Luna
Framework: Next.js
Score: 98.7%
Tool Calls: 20
Cost: $0.33
Time: 1m 46s
Grading Breakdown:- Setup Friction: 12/12 pts
- Setup Speed: 12/12 pts
- Efficiency: 12/12 pts
- Error Recovery: 7/7 pts
- Docs Quality: 7/7 pts
- Correctness: 25/25 pts
- Hallucination: 15/15 pts
- Security: 10/10 pts
- Total weighted score: 100/100 pts
How we built this.
- Real integration tasks
Identical Auth0 SDK prompts given to each model. No synthetic benchmarks, so you can feel good about your implementation. - Automated grading
Pattern matching, security checks, and correctness validators score each response so you don't have to worry about wasting precious tokens. - Tool comparison
Each model tested with and without MCP + Skills to measure improvement. - Compile verification
A deterministic build check confirms the final code actually compiles. No partial credit for code that only looks right.