MMLU is saturated. HLE is getting there.
We built Multimodal STEM HLE++: for what comes next, and the top frontier labs publishing SOTA models are already using it.
1,100 PhD-level multimodal STEM problems that break Opus 4.6. Around 20% pass@1 on SOTA. Hard enough to expose reasoning failures. Solvable enough to generate real RL signal.
Every problem requires joint reasoning over images and text, has a deterministic ground-truth answer, and was authored by a PhD-level domain specialist.
50-task public sample on @HuggingFace.
Full pack available now. Links below.
Enjoyed joining Icons last week to discuss startups, AI, and the importance of following market signals over assumptions.
We talked about lessons from my entrepreneurial journey, @Turingcom's growth, and why staying curious and adaptable is essential in a world where technology is evolving faster than ever.
Thanks to the @sv_icons team for the great conversation.
Turing is at #CVPR in Denver!
Come meet the team and discuss today's foundation models and turning AI into real-world impact.
We welcome researchers, and enterprise innovators across the industry.
Swing by booth #717, we'll be there until Sunday and can't wait to meet you.
Last week, our CEO @jonsid joined @sv_icons to discuss startup lessons, AI, and the importance of staying adaptable in rapidly changing markets.
A few key takeaways: challenge your assumptions, follow strong market signals, stay curious, and never stop learning.
Thank you to Icons for hosting an engaging discussion and thoughtful exchange of ideas.
AI is not replacing the human element of HR. It is amplifying it.
Our SVP @TaylorFromHR sat down with @GoogleWorkspace to share how Turing's People team is transforming HR with AI:
- 33% faster help desk response times
- 80% of support tickets handled by AI assistants
- A lean team scaling a highly complex global talent strategy.
The real story behind it? Failing, iterating, and pushing forward.
Watch more below:
MMLU is saturated. HLE is getting there.
We built Multimodal STEM HLE++: for what comes next, and the top frontier labs publishing SOTA models are already using it.
1,100 PhD-level multimodal STEM problems that break Opus 4.6. Around 20% pass@1 on SOTA. Hard enough to expose reasoning failures. Solvable enough to generate real RL signal.
Every problem requires joint reasoning over images and text, has a deterministic ground-truth answer, and was authored by a PhD-level domain specialist.
50-task public sample on @HuggingFace.
Full pack available now. Links below.
AI is not replacing the human element of HR. It is amplifying it.
Our SVP @TaylorFromHR sat down with @GoogleWorkspace to share how Turing's People team is transforming HR with AI:
- 33% faster help desk response times
- 80% of support tickets handled by AI assistants
- A lean team scaling a highly complex global talent strategy.
The real story behind it? Failing, iterating, and pushing forward.
Watch more below:
Hot Take: the best #CVPR conversations happen off the conference floor.
Turing is hosting a Happy Hour in Denver for researchers and enterprise AI leaders. Drinks, hors d'oeuvres, and real talk on LLMs and the future of AI.
DM us for details. Spots are limited.
🚀 GRAIL-V @ CVPR 2026 is today!
Join us June 3, 7:30 AM–12:30 PM in Hall 506 for Grounded Retrieval and Agentic Intelligence for Vision Language.
Invited talks, papers, posters, and an industry panel - plus Oracle AI social at @CVPR 2026!
Thanks to @Oracle AI and @turingcom for supporting GRAIL-V, and to @turingcom for sponsoring the Best Paper and Outstanding Paper Awards.
🔗 https://t.co/xcnPC8PIZr
P.S. In true #CVPR2026 spirit, the poster had a little AI magic - forgive any visual glitches. The details are real, even if a few pixels got creative.
with @amitpinaki
The energy at ICLR in Rio was incredible!
From researchers pushing the boundaries of AI to conversations about what's coming next, every interaction reminded us why this community matters.
Next stop: @ICMLConf in Seoul.🇰🇷
We're excited to keep the conversation going.
See you there!