Arena
AI evaluation leaderboard platform for comparing LLMs, ranking Gemma 4 as top US open model
Our Take
Arena is basically the scoreboard the AI world actually pays attention to — it's run by LMSYS (the folks behind that massive Chatbot Arena crowdsourcing thing) and they're putting Gemma 4 at the top of the US open-source pile. Look, leaderboards are always debatable, but this one carries more weight than most because it leans on real human comparisons rather than just synthetic benchmarks. If you're tracking what's actually performing in the wild, this is the reference point.
Key Facts
The people behind Arena
Lianmin Zheng
profileKey contributor
Key contributor at LMSYS / Arena. Ph.D. at UC Berkeley (advised by Ion Stoica and Joseph E. Gonzalez). Now member of technical staff at xAI, leading inference team.
Links
Browse by category
Similar products worth knowing

Afterquery
teach machines how experts think

Sakana AI
Japanese AI startup developing hypernetwork methods for instant LoRA 'compilation' - Doc-to-LoRA and Text-to-LoRA genera

Moonshot AI (Kimi)
Chinese AI startup behind the Kimi AI assistant, with Kimi K2.5 being one of the top open models competing with Gemma 4

Manus
AI agent product from Monica AI that fits inside the core agent loop: execute tool → capture result → append to context
Want products like this in your inbox every morning?
Five products. Every morning. Written by someone who actually cares whether they're good or not. Free forever, unsubscribe whenever.