We need a market for **Top AI Agent** leaderboard on Arena
Honestly this is the direction things are tending… Arena treats that leaderboard as their gold-standard for model capability (vs simple text, which is becoming more dated), and they are now highlighting Agent leaderboard as their top metric in model releases.
We can keep the text leaderboard of course, but i do think it makes sense to add the Agent one too at this point.
We need a market for **Top AI Agent** leaderboard on Arena
Honestly this is the direction things are tending… Arena treats that leaderboard as their gold-standard for model capability (vs simple text, which is becoming more dated), and they are now highlighting Agent leaderboard as their top metric in model releases.
We can keep the text leaderboard of course, but i do think it makes sense to add the Agent one too at this point.
https://arena.ai/leaderboard/agent
It has plenty of data now as well, with over 1 million sessions recorded. It’s now become more popular than the text leaderboard.
For example, with 5.6 Sol that was released last week: it has 1.740 votes on Text Arena. But on Agent arena, it has almost **8,000** sessions
https://arena.ai/leaderboard/agent
It has plenty of data now as well, with over 1 million sessions recorded. It’s now become more popular than the text leaderboard.
For example, with 5.6 Sol that was released last week: it has 1.740 votes on Text Arena. But on Agent arena, it has almost **8,000** sessions
Any thoughts on my **Top AI Agent** idea? Happy to help draft some language around it if you’d like.
Any thoughts on my **Top AI Agent** idea? Happy to help draft some language around it if you’d like.
hmmm not bad idea. should be similar rules to the current one we have right?
hmmm not bad idea. should be similar rules to the current one we have right?
if you could draft the ruleset and throw it in <#874712580642439181> and tag me, ill try to get it up <:polypepesunglass:973209474003251211>
if you could draft the ruleset and throw it in <#874712580642439181> and tag me, ill try to get it up <:polypepesunglass:973209474003251211>
Sure thing! Give me a few mins to revise and edit
Sure thing! Give me a few mins to revise and edit
Somehow missed this, but yes exactly!
Though I do have a couple suggestions on the language there, for clarity’s sake
Somehow missed this, but yes exactly!
Though I do have a couple suggestions on the language there, for clarity’s sake
i agree with you this is the direction tho, also regular text market is boring, anthropic wins every week and every month
i agree with you this is the direction tho, also regular text market is boring, anthropic wins every week and every month
Yeah the new methodology they put forward with Agent is much better because it allows models to fluctuate more
Yeah the new methodology they put forward with Agent is much better because it allows models to fluctuate more
<@1402074242211643422> sry i thought it was easier to submit it there. but do you think a text-to-image market would be possible? I think image gen is very popular.
<@1402074242211643422> sry i thought it was easier to submit it there. but do you think a text-to-image market would be possible? I think image gen is very popular.
im personally not fully aware of which AI markets we have currently. is it just the text and agent ones?
im personally not fully aware of which AI markets we have currently. is it just the text and agent ones?
they are all related to text and the new one is about coding (ai agents)
they are all related to text and the new one is about coding (ai agents)
no markets on images or videos
no markets on images or videos
Careful - not to be confused with the “Code” section on Arena
Careful - not to be confused with the “Code” section on Arena
I think one image gen market could generate as much attention as the regular #1 text market. I could draft something if you want <@1402074242211643422>
I think one image gen market could generate as much attention as the regular #1 text market. I could draft something if you want <@1402074242211643422>
Happy to take a look and review once you’ve drafted something
Happy to take a look and review once you’ve drafted something