Translating graphics into tactile charts for blind & low-vision readers requires compliance w/extensive design rules. MIT’s "Tactile Vega-Lite" tool hardwires such guidelines w/i its program, potentially streamlining the process for designers & teachers: https://
bit.ly/4j1EhDM
RESEARCH
-

MIT’s Tactile Vega-Lite Tool Makes Data Charts Accessible
By
–
-
Gemini 2.5 Pro Experimental Achieves Top Performance Benchmarks
By
–
Introducing Gemini 2.5 Pro Experimental! 🎉
— Oriol Vinyals (@OriolVinyalsML) 25 mars 2025
Our newest Gemini model has stellar performance across math and science benchmarks. It’s an incredible model for coding and complex reasoning, and it’s #1 on the @lmarena_ai leaderboard by a drastic 40 ELO margin. Only a handful of… pic.twitter.com/soiAVSD00jIntroducing Gemini 2.5 Pro Experimental! Our newest Gemini model has stellar performance across math and science benchmarks. It’s an incredible model for coding and complex reasoning, and it’s #1 on the @lmarena_ai leaderboard by a drastic 40 ELO margin. Only a handful of
-

Gemini 2.5 Pro outperforms o3-mini-high on reasoning tasks
By
–

Gemini 2.5 Pro evals Looks like it has a much better performance at reasoning tasks than o3-mini-high.
-

Nebula achieves the largest score jump ever on lmarena
By
–
Nebula has a largest score jump ever on lmarena
-

Claude 2.5 Pro Achieves New SOTA Across AI Benchmarks
By
–
2.5 Pro sets new SOTA capabilities across benchmarks, including: — 18.8% on Humanity's Last Exam — 63.8% on SWE-Bench Verified (agentic coding)
— GPQA Diamond and AIME 2025 (STEM)
— Long-context and visual reasoning -

Google Releases Gemini 2.5 Pro, New SOTA AI Model
By
–
BREAKING: Google just dropped Gemini 2.5 Pro, their most intelligent AI model yet.
— Rowan Cheung (@rowancheung) 25 mars 2025
The release debuts at No. 1 on LMArena, and brings new SOTA capabilities across benchmarks for coding, math, science, and more.
More info & demos below: pic.twitter.com/FGpro8gCbwBREAKING: Google just dropped Gemini 2.5 Pro, their most intelligent AI model yet. The release debuts at No. 1 on LMArena, and brings new SOTA capabilities across benchmarks for coding, math, science, and more. More info & demos below:
-

Google Introduces Gemini 2.5 Pro Experimental Model
By
–
Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding. Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on @lmarena_ai leaderboard.
-

Gemini 2.5 Pro thinking model theory questioned
By
–

Got Gemini 2.5 Pro as well, and it disappeared afterwards. It may happen that, in this case, earlier reports on it being a Thinking model may not be true. This would mean that the response came from a different model and not 2.5 Pro.
-
AI Model Performance Evolution and Prompt Adherence Improvements
By
–
Agreed. When it first came out it was pretty impressive with how prompt adherent it was. It’s been surpassed many time since then.
