Ankesh Anand (@ankesh_anand) 's Twitter Profile
Ankesh Anand

@ankesh_anand

Research scientist @googledeepmind (Gemini Thinking), prev phd @milamontreal. Working on RL for reasoning and new capabilities in gemini. Opinions are my own.

ID: 425611240

linkhttps://ankeshanand.com/ calendar_today01-12-2011 06:29:52

992 Tweet

5,5K Followers

650 Following

Ankesh Anand (@ankesh_anand) 's Twitter Profile Photo
Ankesh Anand
@ankesh_anand

7 months ago

The whole surprise over 5.5M$ was because everyone is anchored to Llama3’s compute efficiency. Wenfeng himself said it’s about two generations behind frontier lab numbers. Sonnet costs “tens of millions” of dollars, I hope we release the 2.0 Flash / Flash Thinking numbers as

The whole surprise over 5.5M$ was because everyone is anchored to Llama3’s compute efficiency. Wenfeng himself said it’s about two generations behind frontier lab numbers. Sonnet costs “tens of millions” of dollars, I hope we release the 2.0 Flash / Flash Thinking numbers as
Ankesh Anand (@ankesh_anand) 's Twitter Profile Photo
Ankesh Anand
@ankesh_anand

5 months ago

shoutout to the believers!

shoutout to the believers!
Ankesh Anand (@ankesh_anand) 's Twitter Profile Photo
Ankesh Anand
@ankesh_anand

5 months ago

📈📈📈

📈📈📈
Ankesh Anand (@ankesh_anand) 's Twitter Profile Photo
Ankesh Anand
@ankesh_anand

3 months ago

Here we go! A new 2.5 Pro with all around capability improvements compared to previous versions. - Much better at code editing now, sota on Aider (82.2), try out this model on cursor! - #1 on webdev-arena (surpassing opus 4). - supports budgets now (128 to 32k) - much better at

Here we go! A new 2.5 Pro with all around capability improvements compared to previous versions. - Much better at code editing now, sota on Aider (82.2), try out this model on cursor! - #1 on webdev-arena (surpassing opus 4). - supports budgets now (128 to 32k) - much better at
Kimi.ai (@kimi_moonshot) 's Twitter Profile Photo
Kimi.ai
@kimi_moonshot

a month ago

Kimi K2 tech report just dropped! Quick hits: - MuonClip optimizer: stable + token-efficient pretraining at trillion-parameter scale - 20K+ tools, real & simulated: unlocking scalable agentic data - Joint RL with verifiable + self-critique rubric rewards: alignment that adapts -

Kimi K2 tech report just dropped! Quick hits: - MuonClip optimizer: stable + token-efficient pretraining at trillion-parameter scale - 20K+ tools, real & simulated: unlocking scalable agentic data - Joint RL with verifiable + self-critique rubric rewards: alignment that adapts -