Context: https://huggingface.slack.com/archives/C0A1MNQ8LF3/p1777552593518999

Summary

Improve discoverability of leaderboard-related CLI commands:

• New example in --help: Added hf datasets ls --filter benchmark:official as a third example in the hf datasets leaderboard help text, so users know how to discover available leaderboards.
• Hint: list available leaderboards: After the leaderboard table output, prints a hint suggesting hf datasets ls --filter benchmark:official.
• Hint: get model info: After the leaderboard table output, prints a hint suggesting hf models info <model_id> using the top-ranked model from the results as a concrete example.

Example output

$ hf datasets leaderboard SWE-bench/SWE-bench_Verified --limit 3
RANK MODEL_ID                      VALUE SOURCE    
---- ----------------------------- ----- ----------
   1 deepseek-ai/DeepSeek-V4-Pro    80.6 Model Card
   2 moonshotai/Kimi-K2.6           80.2 Model Card
   3 deepseek-ai/DeepSeek-V4-Flash    79 Model Card
Hint: Use 'hf datasets ls --filter benchmark:official' to list available leaderboards.
Hint: Use 'hf models info deepseek-ai/DeepSeek-V4-Pro' to get details about a model.

Slack Thread