Full deepseek r1. Dont care even if its more expensive than llama 3 405b.
Need speed.
No other r1 providers are providing speed
1 Like
Sambanova is good thing, but 16k context will devalue any model. I want to use aider with 128k coding models, Sambanova has no offer for me ;(
However, I appreciate new Developer plan, at least something.
1 Like
@kollikiran456 R1 is live in the playground and there is a sign up wait list for the API. Given a spin.
-Coby
@sannysanoff I hear you and I am passing your exact quote onto our product teams.
-Coby
It’s Here and available in the playground. There is a signup waitlist for API access.
-Coby
1 Like
Aye… Right now will try it on the playground, waiting for the API to be released.
Make sure you sign up on the wait list.
-Coby
Hi Coby, One comment for it. Even it would be great to have a model with large context window, such the Qwen 2.5-Turbo which capable of 1 million tokens.
That would preprocess the data from the input, and result the answer with a more sophisticated model. I could imagine a flow with two steps to eliminate the iterations with chunks.
One new addition, which could help me for one of my project elements:
Microsoft OmniParser 2.0: