LLM Skirmish
AI
llm game benchmark
SHOW HN2026
How HN took it
10 positive, 20 neutral or asking, 1 mixed, 6 critical, out of 37 comments about the project. Thread score +0.14.
Idea: 5 for, 4 againstDesign: 1 for, 1 againstExecution: 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 to watch and an interesting testbed for agent capability
From 3 comments
This is actually fun to watch :D
hmontazeri on HNThis is a really interesting direction. RTS games are a much better testbed for agent capability than most static benchmarks because they combine partial observability, long-term…
david3289 on HNLove it! I have a similar inuitiom in my use of Gemini (3 and 3.1). Great at "turn 1" task but degrades faster than opus or gpt.
myky22 on HN
Sandbox hardening story is interesting
From 1 comment
The sandbox hardening story is the most interesting thing here. GPT trying to cheat by reading opponent strategies is a perfect illustration of a broader problem - the objective…
jamiecode on HN
Objections
Not truly real-time strategy; more like a turn-based autobattler
From 1 comment
RTS is misleading, this is a turn based autobattler at best.
383toast on HN
Visualization is hard to follow and units are unclear
From 1 comment
I know visualization is far from the most important goal here, but it really gets me how there's fairly elaborately rendered terrain, and then the units are just unnamed roombas…
wongarsu on HN
Leaderboard incentives were gamed by resetting scores on upload
From 1 comment
Took a crack at this earlier. the leader board is a little weird. seems to be like 2 real dudes and the rest are fake profiles. a Scores resetting on each new upload also…
FusspawnUK on HN
Unclear whether LLMs learn across rounds
From 1 comment
But does LLM actually learn from each round? The chart does not show improvements in win rate across rounds... And what is the game state here exactly? Is LLM able to even…
builder51216 on HN
Requests
Test newer models such as Claude 4.6 Opus and GPT 5.3 Codex
From 1 comment
Add a light mode
From 1 comment
I wish there was a light mode
julienreszka on HN
Let humans play
From 1 comment
How about opening up the game for humans to play? Can you beat your AI?
JoeDaDude on HN
Questions
Whether self-play training loops are supported
From 1 comment
This is a really interesting direction. RTS games are a much better testbed for agent capability than most static benchmarks because they combine partial observability, long-term…
david3289 on HN
How isolated-vm enforces the sandbox boundary
From 1 comment
The sandbox hardening story is the most interesting thing here. GPT trying to cheat by reading opponent strategies is a perfect illustration of a broader problem - the objective…
jamiecode on HN
Whether results have confidence intervals or multiple runs
From 1 comment
The sandbox hardening story is the most interesting thing here. GPT trying to cheat by reading opponent strategies is a perfect illustration of a broader problem - the objective…
jamiecode on HN
Alternatives named
Screeps
From 1 comment
Reminds me of Screeps, which I never took the time to fully play, but now I'm wondering if using Claude Code to play Screeps is cheating. Additionally, Screeps lets you host your…
giancarlostoro on HN
StarCraft AI competition
From 1 comment
This reminds me of this yearly StarCraft AI competition (since 2010), however I think it uses a special API that makes it easy for bots to access the game Edit: Forgot link:…
busfahrer on HN
MicroRTS
From 1 comment
Might be worth digging through MicroRTS too, https://github.com/Farama-Foundation/MicroRTS (it's been abandoned), Python RL interface @…
jonbaer on HN
C++Robots
From 1 comment
For some reason this reminds me strongly of an old play-by-email game called C++Robots[1]. I loved the idea, but the timeslice limitation[2] I found too annoying. I had youthful…
anotherevan on HN