⁵⁄₉
@fiveoutofnine
Followers
10K
Following
26K
Media
1K
Statuses
3K
results: https://t.co/9AsK05qIYD if you want to try submitting/beating it: https://t.co/GhXfUUv7OC
peval.io
Solve Sudoku puzzles of varying sizes and difficulty.
0
0
3
> o3 is the only model that solves any of the 9x9 Sudokus gpt-oss-120b is also able to solve 9×9s (1.4%). the only other model on peval that solved any 9×9s is GPT 5
1
0
2
NEW JERSEY: Get to know MIKIE SHERRILL: —Voted to RAISE taxes by $3,700 —Voted to give ILLEGALS taxpayer-funded healthcare —Voted to CUT the child tax credit And Sherrill has NO idea what to do as Governor. 🤯 Vote AGAINST her on Nov. 4th.
88
192
532
gpt-oss-120b is so good ties Gemini Pro 2.5 here and is 98.9% cheaper
Following our Sudoku-based reasoning benchmark announcement, we've been evaluating the latest models to track improvements in their reasoning capabilities. Today, we’re launching the Sudoku-Bench Leaderboard: https://t.co/uSreGcB7NQ New technical report: https://t.co/1715s0UNQl
1
0
8
BREAKING: Gavin Newsom officially considers 2028 presidential campaign. 22% chance he’s the next POTUS.
799
243
3K
feel like it's the sort of thing that feels hard, then you just go do it, and then it's like k that was fine
1
0
4
I actually think most people would instead be surprised at how easy these distances are because the "idea" of it is probably much harder their minds e.g. anyone healthy could walk a marathon (10 hours of waking) on any given day if they had to
every man should run a marathon, 50 miles, 100 miles, once in their life marathon takes a few hours 50 miles takes 9-12 hours 100 miles takes 24-36 hours not that much time of your life, and it resets your perspective on pain forever
6
0
24
Delighted to announce @0xren_cf has been promoted to General Partner at @ElectricCapital! Ren started as an engineer at Electric & has grown into a phenomenal investor and unique thinker. Founders are lucky to work with him. We are lucky to have him at Electric. Link below 👇
99
26
461
The scariest stories this October aren’t fiction—they’re funded. Read the new Capital Research magazine issue on our website!
4
14
109
reminds me of the time my on-chain chess NFT had 0 activity then 8 mints came through in the same tx from @z0age
Not many people know this, but there's been a full game of smart contract vs smart contract chess on mainnet ~2 years ago Seeing @z0age's contract beat mine was like the happiest moment of that month for me
0
0
5
so surprised to see 18 submissions come in at the same time lmao gemini 2.5 pro is the best so far:
ok so i ran some prompts on the Sudoku Prompting Competition from Peval (@fiveoutofnine) - set up my API keys, drafted a few prompt styles, and spam-submitted runs to test behaviors. - tried 4 prompt types: 1)be creative, 2)threat to replace it, 3)you are a genius and 4)standard
1
0
9
Healthcare DEFLATION continues: Family of 4 paid $530 in Nov 24 Family of 4 will pay $505 in Nov 25 5% reduction Individual (<55) paid $160 in Nov 24 Individual (<55) will pay $150 in Nov 25 6% reduction
10
12
190
the idea with this is anyone can fund the "research" of prompting on a task they'd like to see progress on
prize pools are live on peval! compete to win USDC or help incentivize prompting by contributing to pools
0
0
15
win $16 for getting #1 on the sudoku competition (will be paid out ~nov. 7, after the competition ends): https://t.co/GhXfUUv7OC
peval.io
Solve Sudoku puzzles of varying sizes and difficulty.
1
0
5
prize pools are live on peval! compete to win USDC or help incentivize prompting by contributing to pools
1
0
13
GraniteShares announced distribution rates for the YieldBOOST™ COIN ETF (COYY) as of October 30, 2025 $COYY COYY: Standardized performance: Since Inception (7/29/2025): -15.20%
3
10
75
I asked plotchy to be an early tester for peval, and then he full-scored it 1st try and sent me this like 3 hours later 👹
Multiplying numbers used to be an LLM gotcha but now is nearly solved. For this competition I ran a grid search over all mainstream models and gpt-oss-120b stood out scoring ~93% correct on 18digit*18digit multiplication! Insane! eg: 364826485628193748 * 492816485726395817 = ...
2
0
17
even more true for multiplication https://t.co/YELyXI0XPd
interestingly, 𝚐𝚙𝚝-𝚘𝚜𝚜-𝟷𝟸𝟶𝚋 was the best model w/ a perfect score, and all non-OpenAI models struggled to score above 0.58
0
0
3