Nous Research
@NousResearch
World-class open source models https://discord.gg/nousresearch
You might like
Today we open source Nomos 1. At just 30B parameters, it scores 87/120 on this year’s Putnam, one of the world’s most prestigious math competitions. This score would rank #2/3988 in 2024 and marks our first step with @hillclimbai towards creating a SOTA AI mathematician.
I am once again reaching out to you all to apply @ nous for the post training team to work with me and our team on advancing several core capabilities of our models! If you want to work on: - Creativity, Roleplaying, and Simulation - STEM - Math - Code and Code Agents -…
We are hiring highly skilled front end/ux developers, MLE's focused on RL, Training Infrastructure, MLOps, and Pretraining, and some other positions! nousresearch.com/careers
nousresearch.com
Careers - NOUS RESEARCH
CAREERS OUR MISSION is to create and democratize access to the world’s best intelligence. Powerful models should be in the hands of the many rather than the privileged few. To get there, we’ve...
We just released our first Hermes model trained entirely with Distro on Psyche, Hermes 4.3 on ByteDance Seed 36B! Outcome was actually better than the centralized comparison run, and even brought it to top spot on the RefusalBench leaderboard, weights on HF! Check it out:
Presenting Hermes 4.3 on ByteDance Seed 36B, the latest update to our flagship Hermes series of models. This model offers roughly equivalent performance to Hermes 4 70B at half the model size, and was post-trained entirely on the Psyche network secured by @Solana.
Want to learn more about how Psyche outperformed traditional centralized post-training? Join the team for an Office Hours session in 2 hours → discord.gg/nousresearch
Presenting Hermes 4.3 on ByteDance Seed 36B, the latest update to our flagship Hermes series of models. This model offers roughly equivalent performance to Hermes 4 70B at half the model size, and was post-trained entirely on the Psyche network secured by @Solana.
Ok -- so @NousResearch seriously buried the lead here. For this most recent model run, they've validated that their decentralized training harness, Psyche, outperformed centralized training. Why is this massive? It means with enough nodes, people could take on the biggest labs.
In addition to the new high-limit subscription, we have rolled out two new updates: ⁃ Starting today you can use Nous Chat without an account, anonymously and for free ⁃ The Nous API now supports payment for inference with USDC on @solana, facilitated by @CoinbaseDev's x402
Get a free month of Nous Chat to celebrate Cyber Monday with code CYBER2025 (valid for one day only)! Access Hermes 3 & 4 as well as a variety of frontier models with high usage limits and deeply configurable chat settings
United States Trends
- 1. Merry Christmas Eve 51.4K posts
- 2. Spurs 50.6K posts
- 3. #Pluribus 21.7K posts
- 4. Rockets 25.1K posts
- 5. hudson 150K posts
- 6. Cooper Flagg 12.8K posts
- 7. UNLV 2,616 posts
- 8. Chet 10.6K posts
- 9. SKOL 1,757 posts
- 10. Ime Udoka N/A
- 11. Dillon Brooks 5,070 posts
- 12. Rosetta Stone N/A
- 13. Mavs 6,398 posts
- 14. #PorVida 1,830 posts
- 15. #VegasBorn N/A
- 16. Yellow 58.9K posts
- 17. Randle 2,702 posts
- 18. connor 155K posts
- 19. Kawhi Leonard 1,264 posts
- 20. #ClipperNation N/A
You might like
-
Eric Hartford
@QuixiAI -
lmarena.ai
@arena -
clem 🤗
@ClementDelangue -
Alignment Lab AI
@alignment_lab -
Yam Peleg
@Yampeleg -
Teknium (e/λ)
@Teknium -
Jerry Liu
@jerryjliu0 -
Abhi Venigalla
@ml_hardware -
Gradio
@Gradio -
anton
@abacaj -
Jon Durbin
@jon_durbin -
Sam Havens
@sam_havens -
Philip Turner
@philipturnerar -
Yao Fu
@Francis_YAO_ -
Philipp Schmid
@_philschmid
Something went wrong.
Something went wrong.