Sri Kundurthy

@stanford

[Sri Kundurthy @srkvatsa]

Heading to ICML next week! 🇰🇷

Excited to share 3 accepted works:

- SpreadsheetArena
- On-Policy Expert Corrections
- SWE-Bench Pro

Would love to talk about agent evals, post-training, and data-constrained scaling laws.

Looking to connect with future collaborators @stanford!

[Zach Kirshner @Zachkirshner]

day 1 in the books

dm me if you’re around and want to meet up!
#icml #icml26

[Sri Kundurthy @srkvatsa]

I’ll be presenting our work on on-policy expert corrections (OEC) at ICML next week!

grateful for the team led by @NiklasLauffer, with Xiang Deng, @Bckenstler and @_jeffda

[Scale Labs @ScaleAILabs]

Imitation Learning for Multi-turn LM Agents via On-policy Expert Corrections
https://t.co/0TGQpXpEFb

[Sri Kundurthy @srkvatsa]

excited to present our paper on Spreadsheet Arena, co-led with @claranahhh and the @MeridianAgent team, at ICML next week!

[Spreadsheet Arena @sheetarena]

Spreadsheets have entered the arena! ⚔️
Announcing Spreadsheet Arena, the first research platform for human preference rankings on LLM-generated spreadsheets.

The results? @AnthropicAI Claude Opus is on top, but the gap is tighter than you’d think.

w/ @LTIatCMU, @Cornell, and @scale_ai. 🧵

[Scale Labs @ScaleAILabs]

SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?
https://t.co/KZzZfD5sL4

[Calvin Zhang @calvincbzhang]

New paper from @scale_AI & @MeridianAgent: SpreadsheetArena 📄

We evaluated 16 LLMs on end-to-end spreadsheet generation via 4,300+ blind pairwise votes.

Crucially, we move beyond scalar Elo ratings to decompose the latent preference signal into functional, structural, and stylistic components. 🧵

[Scale AI @scale_AI]

Honored to see SWE-Bench Pro recognized as the new standard for frontier coding evals. 💜

We built it to address saturation and contamination in earlier benchmarks — raising the bar with a more rigorous measure of agents' problem-solving capabilities and a clearer view of real-world software engineering progress.

[Scale AI @scale_AI]

HLE, SWE-Bench Pro + MCP Atlas:

https://t.co/UkZt6bzN6P

[Spreadsheet Arena @sheetarena]

Sonnet 4.6 is now on Spreadsheet Arena!

How well can it model Anthropic's Series G?

[Spreadsheet Arena @sheetarena]

Spreadsheets have entered the arena! ⚔️
Announcing Spreadsheet Arena, the first research platform for human preference rankings on LLM-generated spreadsheets.

The results? @AnthropicAI Claude Opus is on top, but the gap is tighter than you’d think.

w/ @LTIatCMU, @Cornell, and @scale_ai. 🧵

[TechCrunch @TechCrunch]

Meridian raises $17 million to remake the agentic spreadsheet https://t.co/1cK2AIJU0K

[Sam Altman @sama]

GPT-5.3-Codex is here!

*Best coding performance (57% SWE-Bench Pro, 76% TerminalBench 2.0, 64% OSWorld).
*Mid-task steerability and live updates during tasks.
*Faster! Less than half the tokens of 5.2-Codex for same tasks, and >25% faster per token!
*Good computer use.

[Alexandr Wang @alexandr_wang]

new @scale_AI era loaded 🫡

- q4 was our biggest quarter ever
- US gov business growing faster than ever
- data business = profitable
- multiple 9 figure enterprise + gov deals

[Sam Altman @sama]

Performance is strong across the board: 55.6% on SWE-Bench Pro, 52.9% or ARC-AGI-2, 40.3% on Frontier Math.

[Bing Liu @vbingliu]

🚀 Introducing SWE-Bench Pro — a new benchmark to evaluate LLM coding agents on real, enterprise-grade software engineering tasks.

This is the next step beyond SWE-Bench: harder, contamination-resistant, and closer to real-world repos.

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains

'We introduce Rubrics as Rewards (RaR), a framework that uses structured, checklist-style rubrics as interpretable reward signals for on-policy training with GRPO. Our best RaR method yields up to a relative improvement on HealthBench-1k compared to simple Likert-based approaches, while matching or surpassing the performance of reward signals derived from expert-written references."

[AI at Meta @AIatMeta]

Today we’re releasing Code Llama 70B: a new, more performant version of our LLM for code generation — available under the same license as previous Code Llama models.

Download the models ➡️ https://t.co/GApdj5PW83
• CodeLlama-70B
• CodeLlama-70B-Python
• CodeLlama-70B-Instruct

[Jim Fan @DrJimFan]

What if we set GPT-4 free in Minecraft? ⛏️

I’m excited to announce Voyager, the first lifelong learning agent that plays Minecraft purely in-context. Voyager continuously improves itself by writing, refining, committing, and retrieving *code* from a skill library.

GPT-4 unlocks a new paradigm: ��training” is code execution rather than gradient descent. “Trained model” is a codebase of skills that Voyager iteratively composes, rather than matrices of floats. We are pushing no-gradient architecture to its limit.

Voyager rapidly becomes a seasoned explorer. In Minecraft, it obtains 3.3× more unique items, travels 2.3× longer distances, and unlocks key tech tree milestones up to 15.3× faster than prior methods.

We open-source everything. Let generalist agents emerge in Minecraft! Welcome you all to try today: https://t.co/1d3YocozsI
Paper: https://t.co/JcWEasgtyI
Code: https://t.co/KsvVf7rcl0

Deep dive with me: 🧵

[LAION @laion_ai]

We launche a petition to democratize AI research by establishing an international, publicly funded supercomputing facility equipped with 100,000 state-of-the-art AI accelerators to train open source foundation models.
https://t.co/jbkQnlEHYH
https://t.co/agmh1Ptjky