How come deepseek v2 is so low in the leaderboard?

#837
by ZeroWw - opened

The logicAL problem solving skills of deepseek v2 (normal and coder) are almost on par with claude and the best I have seen so far. Absolutely uncomparable with qwen2 and others.
Gemini pro seems to ge at the same level or slightly more than deepseek and less than claude.

oh wait.. it's not even in the leaderboard! that was the old version. weird.

Open LLM Leaderboard org

Feel free to submit it if you want it to be evaluated!

clefourrier changed discussion status to closed

Sign up or log in to comment