You can basically work with agents to write markdown skills and put it on a cron job and do almost every useful type of knowledge work now

Grok 4.7 takes the TOP 3 spots on VulcanBench Frontier v4, outperforming Fable 5.1, Opus 5.5, GPT-6 Astra, GPT-6.1 Sol and other leading frontier models • Grok 4.7 xHigh - 93.15 • Grok 4.7 High - 92.71 • Grok 4.7 Medium - 92.30 VulcanBench Frontier v4 tests whether AI models can handle difficult repository-level software engineering work rather than simply generate isolated code snippets The benchmark contains 23 behavioural-reconstruction tasks where models have to rebuild legacy software to match its actual behaviour, with functional correctness tested through hidden tests alongside security, lint/complexity and code quality And the task results are just as impressive: • Grok 4.7 xHigh - 23/23 passed • Grok 4.7 High - 22/23 passed • Grok 4.7 Medium - 21/23 passed Grok 4.7 took all #1, #2 and #3 spots

Here's a fun benchmark we ran: 8 decision models playing Tetris. In our tests @perplexity_ai Decider came consistently at the top !! ( cc @AravSrinivas ) you can try em out here - https://app.routerplus.com/playground/tetris-royale

@julianharris There's something wet-sponge-feeling about the qwen fine tunes that I can't explain compared to Jev

“We find that on medium-length, well-defined accounting tasks, frontier AI models are now faster and more accurate than junior accountants, even the best one in our study.” Eighteen months ago they scored well below human accountants Good discussion here: https://www.mercor.com/blog/human-baselines-for-benchmarks-ai-now-outperforms-junior-accountants/

'More striking is how fast AI took the lead. Just eighteen months ago, the best AI models fell short of the average accountant’s ~37% score. Today, models ace those same tasks.' 'These results are provocative. So much so that we considered not publishing them for fear of misinterpretation. But we think transparency about the findings matters as people and institutions prepare for rapidly advancing AI.'

@Jabaluck @DAcemogluMIT The meta-benchmark of being able to pass ARC-AGI-(n+1) immediately upon release. Though presumably that won't last forever.

STOP making "Jevbench"es, stop asking for public benchmarks, they completely miss the point of Jev and you won't believe how easy it is to game every benchmark you hold dear This is @CompleteSkeptic's bitterest lesson of all: picking the right task beats everything

ARC Prize 2026 - ARC-AGI-3 Progress Prize - 9 day left $37,500 in prizes are being awarded on *Sept 30th* to the top open source solutions Current standings: 1. Tufa Labs 2. Lord Han Solo 3. NVARC3 Which top 3 will claim the open source prize?