Chinese AI Tops Hugging Face's Revamped Chatbot Leaderboard

[erek]

Fascinating.. also DeepSeek Coder outperforms GPT-4?

"Alibaba's Qwen models dominated Hugging Face's latest LLM leaderboard, securing three top-ten spots. The new benchmark, launched Thursday, tests open-source models on tougher criteria including long-context reasoning and complex math. Meta's Llama3-70B also ranked highly, but several Chinese models outperformed Western counterparts. (Closed-source AIs like ChatGPT were excluded.) The leaderboard replaces an earlier version deemed too easy to game."

Source: https://slashdot.org/story/24/06/27...ps-hugging-faces-revamped-chatbot-leaderboard

 

[LukeTbk]

Microsoft relatively small model score well here, but excluding the new Claude, Geminy and GPT can make it hard to have a good view of how good they are.

DeepSeek seem to beat the older Claude 3 and GPT-4 which is really good, does not include test against 3.5 or gpt 4-o.

can by tried easily here (much slower but still fast enough):
https://chat.deepseek.com/coder

 

[Thatguybil]

It’s been exciting to see the advancements in the health terminology leaderboards.

Twice the public models have leapfrogged our internal models… of which I have no endless fun pointing out to our data scientists…

 

[erek]

It’s been exciting to see the advancements in the health terminology leaderboards.

Twice the public models have leapfrogged our internal models… of which I have no endless fun pointing out to our data scientists…

 

[CookieFactory]

BABA most undervalued AI play IMO. Not to mention how good the rest of their business is looking.