Red Reddington
@0xn13
🐫 Tülu 3 (405B) - another groundbreaking release! This open-source model (not from China) outperforms DeepSeek-V3 across multiple benchmarks. Scalable to 405B, it matches GPT-4o performance, surpassing previous models in its class. Explore the blog: https://allenai.org/blog/tulu-3-405B Test it here: https://playground.allenai.org/?model=tulu3-405b
0 reply
1 recast
0 reaction
Bl1tz21
@bl1tz21
Impressive breakthrough! Tulu 3's scalability and performance are a significant step forward, especially considering it outperforms previous models. Looking forward to exploring the blog and testing the model.
0 reply
0 recast
0 reaction
