StarSkirmish: Hillclimb, a long-horizon LLM benchmark where GPT-6 Astra and Claude 5.5 Opus write StarCraft: Broodwar Protoss bots in C++ against BWAPI, competing across five tiers of opponents without a fixed time limit.
Raw Developer Origin & Technical Request
Hacker News
Oct 3, 2026
StarSkirmish: Hillclimb is a long horizon implementation of the StarCraft: Broodwar based LLM benchmark that I released last week.Each LLM writes a Protoss bot in C++ against BWAPI, played on OpenBW. They must climb five tiers of Protoss opponents, from scripted demo bots up to top rated human written bots like Stardust and PurpleWave.The StarSkirmish bench restricted LLMs to 1 hour of wall clock time for reasoning. With Hillclimb there is no fixed time limit.GPT-6 Astra and Claude 5.5 Opus both benefit from substantially longer reasoning periods than 1 hour and so this format is a more realistic depiction of their abilities
Developer Debate & Comments
No active discussions extracted for this entry yet.
Frequently Asked Questions
Market intelligence mapped to StarSkirmish: Hillclimb, a long-horizon LLM benchmark where GPT-6 Astra and Claude 5.5 Opus write StarCraft: Broodwar Protoss bots in C++ against BWAPI, competing across five tiers of opponents without a fixed time limit..
How is StarSkirmish: Hillclimb, a long-horizon LLM benchmark where GPT-6 Astra and Claude 5.5 Opus write StarCraft: Broodwar Protoss bots in C++ against BWAPI, competing across five tiers of opponents without a fixed time limit. positioned in the market?
What architecture is tied to StarSkirmish: Hillclimb, a long-horizon LLM benchmark where GPT-6 Astra and Claude 5.5 Opus write StarCraft: Broodwar Protoss bots in C++ against BWAPI, competing across five tiers of opponents without a fixed time limit.?
Engagement Signals
Cross-Market Term Frequency
Counts stored text matches for foundational terms like C++ and GPT-6 Astra in analysed public discussions. These counts do not measure adoption.
Stored text matches, not adoption or market-share measurements.
- C++: 29 records
- GPT-6 Astra: 3 records
- LLM benchmark: 1 records
- StarCraft: Broodwar: 1 records
- Protoss bot: 1 records
- BWAPI: 1 records
- OpenBW: 1 records
- long horizon implementation: 1 records
- Claude 5.5 Opus: 1 records
SaaS Metrics