DataChiBenchmark
Pricing
DataChi

The benchmark for real-world AI. Built and hosted in the EU.

Product
AI GatewayLLM APIEU Sovereign GatewayObservabilityCompare modelsModel race
Benchmarks
LeaderboardModelsTasksMethodologyBest AI for…
Resources
DocumentationAPI referenceBlogPricing
Company
AboutEU AICloud ActPrivacyTermsContact
Newsletter

Get notified when new models are added to the leaderboard.

All systems normalEU sovereign
© 2026 DataChi · Made in Europe with ♥
PrivacyTermsCompliance
Tasks

The 7 things businesses actually ask LLMs to do.

Every model is scored on these categories. Open one to see the prompts, the judges, and the leaderboard for that task.

84 prompts
Code generation

142 models tested · updated daily

Sample prompt
Write a TypeScript REST endpoint that validates a JWT and returns user data.
Top: Claude Sonnet 4Open
52 prompts
Reasoning & logic

142 models tested · updated daily

Sample prompt
Solve this multi-step math problem with a step-by-step explanation…
Top: o3Open
38 prompts
Email writing

142 models tested · updated daily

Sample prompt
Reply to this support ticket professionally and suggest a resolution.
Top: GPT-4oOpen
47 prompts
Customer support

142 models tested · updated daily

Sample prompt
A user is angry about a missed delivery — write a response that de-escalates.
Top: Claude Sonnet 4Open
31 prompts
Legal review

142 models tested · updated daily

Sample prompt
Identify the top 3 risks in this MSA indemnification clause.
Top: GPT-4oOpen
29 prompts
Financial analysis

142 models tested · updated daily

Sample prompt
Compute the IRR for this cash-flow series and summarise the investment thesis.
Top: Claude Sonnet 4Open
35 prompts
Summarisation

142 models tested · updated daily

Sample prompt
Summarise this 40-page meeting transcript in 5 executive bullets.
Top: Gemini 1.5 ProOpen

Don't see your task?

Upload your own prompts. We'll benchmark them across 50+ models and give you a private leaderboard.