Julius Research

This account is verified because it's an official organization on X. Learn more

@juliusai

Applied AI research lab based in San Francisco 🌁

More about us: julius-research.com and julius.ai

San Francisco julius.ai/?utm_source=tw… Joined November 2022

148 Following 21K Followers 767 Posts

Latest posts 5

Julius Research @juliusai 9h

Gemini 4 Carbon is the only model outside of OpenAI and Anthropic in the top 7 on DeepFinance bench

rahul Verified account @0interestrates 9h

Gemini 4 Carbon scores 56.6% on ultra long horizon finance tasks and takes #2 spot on DeepFinance bench

nice work by the GDM team

View quoted post on X

Julius Research @juliusai 3d

GPT-6 Astra at max reasoning is the new leader for ultra long-horizon finance work on DeepFinance leaderboard🥇. After releasing DeepFinance bench, we wanted to see whether more thinking helps on ultra long-horizon finance tasks or if it just added “token-noise”. Turns out, it helps. GPT-6 Astra on max thinking scores 57.6%, exceeding Claude Opus 5.5 (54.8%) and Claude Sonnet 5.5 (54.6 %)

see leaderboard and blog here: juliusresearch.ai/blog/deep-fina…

0:05

rahul Verified account @0interestrates 5d

Today we're releasing DeepFinance, the first ultra-long horizon benchmark for high finance work.

A set of 21 problems that reflect work at top PE and IB firms, DeepFinance is built using real deal data.

Problems are constructed by experts with backgrounds from top firms like Evercore, JPM and UBS, and institutions like HBS and Wharton. Problems represent work that would take human professionals 20-40 hours to complete.

Can AI really diligence a $100M+ deal? Finance benchmarks today are limited to evaluating agents on QA on financial docs or grading the numerical correctness of simple financial models. However, real world high finance work is full of assumptions and conflicting evidence. We built DeepFinance to realistically evaluate agents on those end-to-end finance workflows.

DeepFinance leaderboard: juliusresearch.ai/deep-finance

View quoted post on X

Julius Research @juliusai 5d

Long-horizon trajectory data is critical for expanding the production possibilities frontier of the economy.

Garry Tan Verified account @garrytan 5d

This is the first long horizon benchmark for PE/IB work

Solving high finance is on the path to AGI
View quoted post on X

Julius Research @juliusai Sep 28

Claude Sonnet 5.5 is now live in Julius.

We're seeing it perform stronger in coding tasks, with clearer writing, and faster output than Sonnet 5 in Julius. Put it to work on your next analysis, presentation, or website, and let us know what you think.

Julius Research @juliusai Sep 22

it's equally important to UNSHIP features when they’re no longer necessary

rahul Verified account @0interestrates Sep 9

un-introducing notebooks in @juliusai

View quoted post on X