TetrisBench
AI
llm benchmark
SHOW HN2026
111points
41comments
#611of 4,366 in 2026
Top 14%of 2026 by points
How HN took it
CuriousMostly questions and neutral comments
0 positive, 13 neutral or asking, 1 mixed, 6 critical, out of 20 comments about the project. Thread score -0.30.
Idea: 0 for, 4 againstExecution: 0 for, 1 againstPrivacy: 0 for, 1 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Fun concept and useful feedback for the game mechanics
From 2 comments
Very cool! I am a good Tetris player (in the top 10% of players) and wanted to give brick yeeting against an LLM a spin. Some feedback: - Knowing the scoring system is helpful…
bubblesorting on HNStackflow looks nice! I'm a Balatro fan and I didn't know about this variant. By the way, kudos on your feedback. If I was OP, I would've been honored to get that type of…
kinduff on HN
Objections
Results don't show useful transfer to real tasks
From 2 comments
LLMs playing Tetris feels like testing a calculator's ability to write poetry. Interesting as a curiosity, but the results don't transfer to the tasks where these models actually…
augusteo on HN... and what does this prove? what can you decide to use one LLM to solve over another based on this tetrisbench besides play tetris?
segmondy on HN
Comparing against Stockfish or humans is not a fair test
From 2 comments
Comparing against stockfish isn't fair. That's comparing against enormous amounts of compute spent experimenting with strategies, training neutral nets, etc. It will lose so…
gpm on HNIt's not. Maybe if you used old versions of stockfish that predate the neural net methods used by current versions, because otherwise you'd be comparing the hand-rolled (by an…
mikkupikku on HN
Tetris is solvable with simple algorithms, so LLMs add little
From 2 comments
Guys, I don't know how to tell you but... Tetris can web solved without LLM...
p0w3n3d on HNWhy run something in a few CPU cyles on a 40 year old home computer when you can do the same (but worse) on a billion-dollar GPU cluster?
brookman64k on HN
Latency makes LLM real-time play questionable
From 2 comments
How does it deal with latency? Afaict remote LLMs need seconds to process, but Tetris can move much faster..
dakom on HNLLMs playing Tetris feels like testing a calculator's ability to write poetry. Interesting as a curiosity, but the results don't transfer to the tasks where these models actually…
augusteo on HN
Requests
Fix rotation, RNG and DAS to match standard Tetris
From 2 comments
Very cool! I am a good Tetris player (in the top 10% of players) and wanted to give brick yeeting against an LLM a spin. Some feedback: - Knowing the scoring system is helpful…
bubblesorting on HNThe worst thing is that the delayed auto shift is slightly off and it messes my finesse. (I used to play competitive tetris as well, but between getting older -> worse reflexes…
qsort on HN
Show scoring system and allow remappable keys
From 1 comment
Very cool! I am a good Tetris player (in the top 10% of players) and wanted to give brick yeeting against an LLM a spin. Some feedback: - Knowing the scoring system is helpful…
bubblesorting on HN
Explain how the models receive game state
From 1 comment
Interesting but frustratingly vague on details. How exactly are the models playing? Is it using some kind of PGN equivalent in Tetris that represents a on-going game, passing an…
vunderba on HN
Questions
How are models given the board state and how is latency handled?
From 3 comments
Interesting but frustratingly vague on details. How exactly are the models playing? Is it using some kind of PGN equivalent in Tetris that represents a on-going game, passing an…
vunderba on HNHow does it deal with latency? Afaict remote LLMs need seconds to process, but Tetris can move much faster..
dakom on HNLLMs playing Tetris feels like testing a calculator's ability to write poetry. Interesting as a curiosity, but the results don't transfer to the tasks where these models actually…
augusteo on HN
Why Gemini Flash beats Opus and what drives the difference
From 2 comments
Is there a tl;dr on why this is? Does it just make faster decisions?
indigodaddy on HNmy unvalidated theory is that this comes down to the coding model’s training objective: Tetris is fundamentally an optimization problem with delayed rewards. Some models seem to…
ykhli on HN
Alternatives named
Stackflow
From 1 comment
Hello fellow Tetris nerd with a -sort username :) On the topic of reflexes decaying (I'm getting there, in my late 30s): Have you played Stackflow? It's a number go up roguelite…
bubblesorting on HN
Stacky
From 1 comment
cool to hear about more randomizers. I am not a great tetris player but I absolutely love the game. I put NES, 7-bag and my own take on randomizer (crap) in my online tetris-like…
Keyframe on HN
Brian Moore's AI clocks
From 1 comment
Thanks for the clarification! Kind of reminds me of the Brian Moore's AI clocks which uses several LLMs to continuously generate HTML/CSS to create an analog clock for…
vunderba on HN