AI
Brood War Bench pits AI coding agents against each other in StarCraft
Report details model-specific strategies and failures
1
Sep 19 10:44 AM
· 8d ago · 1 article · 1 post · 2 sources · development 1 of 3
The write-up describes Codex's effective but shallow Probe-harassment tactics, its subagents' poor coordination, Grok 4.6's long reasoning traces with almost no issued commands, and Claude Fable's comparatively ambitious tech-climbing and economy-building, while stressing that a beginner human doing a photon rush would beat every model tested.
bw.swerdlow.dev
“Codex's strongest recurring idea was disruption.”
benswerd
benswerd Creator of Brood War Bench OpenAI Codex AI agent tested in the benchmark Grok 4.6 (xAI) AI agent tested in the benchmark Claude Fable (Anthropic) AI agent tested in the benchmark
The whole story articles posts
the bright band is this development · numbered dots are the others · click one to jump
Peak 9 pieces in two hours at Sep 19, 3 PM; 43 pieces over 8 days (2 articles · 6 posts · 35 comments)
Sep 19, 9 AM — 3 pieces · 2 articles · 1 post — Newswires 2, Hacker News 1 Sep 19, 11 AM — quiet Sep 19, 1 PM — quiet Sep 19, 3 PM — 9 pieces · 9 comments — Hacker News 9 Sep 19, 5 PM — 3 pieces · 1 post · 2 comments — Hacker News 2, Mastodon 1 Sep 19, 7 PM — 1 piece · 1 comment — Hacker News 1 Sep 19, 9 PM — 4 pieces · 4 comments — Hacker News 4 Sep 19, 11 PM — 1 piece · 1 post — Mastodon 1 Sep 20, 1 AM — 4 pieces · 4 comments — Hacker News 4 Sep 20, 3 AM — 7 pieces · 2 posts · 5 comments — Hacker News 5, Mastodon 2 Sep 20, 5 AM — 3 pieces · 1 post · 2 comments — Hacker News 2, Mastodon 1 Sep 20, 7 AM — 2 pieces · 2 comments — Hacker News 2 Sep 20, 9 AM — 2 pieces · 2 comments — Hacker News 2 Sep 20, 11 AM — 2 pieces · 2 comments — Hacker News 2 Sep 20, 1 PM — 1 piece · 1 comment — Hacker News 1 Sep 20, 3 PM — 1 piece · 1 comment — Hacker News 1 Sep 20, 5 PM — quiet Sep 20, 7 PM — quiet Sep 20, 9 PM — quiet Sep 20, 11 PM — quiet Sep 21, 1 AM — quiet Sep 21, 3 AM — quiet Sep 21, 5 AM — quiet Sep 21, 7 AM — quiet Sep 21, 9 AM — quiet Sep 21, 11 AM — quiet Sep 21, 1 PM — quiet Sep 21, 3 PM — quiet Sep 21, 5 PM — quiet Sep 21, 7 PM — quiet Sep 21, 9 PM — quiet Sep 21, 11 PM — quiet Sep 22, 1 AM — quiet Sep 22, 3 AM — quiet Sep 22, 5 AM — quiet Sep 22, 7 AM — quiet Sep 22, 9 AM — quiet Sep 22, 11 AM — quiet Sep 22, 1 PM — quiet Sep 22, 3 PM — quiet Sep 22, 5 PM — quiet Sep 22, 7 PM — quiet Sep 22, 9 PM — quiet Sep 22, 11 PM — quiet Sep 23, 1 AM — quiet Sep 23, 3 AM — quiet Sep 23, 5 AM — quiet Sep 23, 7 AM — quiet Sep 23, 9 AM — quiet Sep 23, 11 AM — quiet Sep 23, 1 PM — quiet Sep 23, 3 PM — quiet Sep 23, 5 PM — quiet Sep 23, 7 PM — quiet Sep 23, 9 PM — quiet Sep 23, 11 PM — quiet Sep 24, 1 AM — quiet Sep 24, 3 AM — quiet Sep 24, 5 AM — quiet Sep 24, 7 AM — quiet Sep 24, 9 AM — quiet Sep 24, 11 AM — quiet Sep 24, 1 PM — quiet Sep 24, 3 PM — quiet Sep 24, 5 PM — quiet Sep 24, 7 PM — quiet Sep 24, 9 PM — quiet Sep 24, 11 PM — quiet Sep 25, 1 AM — quiet Sep 25, 3 AM — quiet Sep 25, 5 AM — quiet Sep 25, 7 AM — quiet Sep 25, 9 AM — quiet Sep 25, 11 AM — quiet Sep 25, 1 PM — quiet Sep 25, 3 PM — quiet Sep 25, 5 PM — quiet Sep 25, 7 PM — quiet Sep 25, 9 PM — quiet Sep 25, 11 PM — quiet Yesterday, 1 AM — quiet Yesterday, 3 AM — quiet Yesterday, 5 AM — quiet Yesterday, 7 AM — quiet Yesterday, 9 AM — quiet Yesterday, 11 AM — quiet Yesterday, 1 PM — quiet Yesterday, 3 PM — quiet Yesterday, 5 PM — quiet Yesterday, 7 PM — quiet Yesterday, 9 PM — quiet Yesterday, 11 PM — quiet Today, 1 AM — quiet Today, 3 AM — quiet Today, 5 AM — quiet Today, 7 AM — quiet Today, 9 AM — quiet Today, 11 AM — quiet
1 2–3
Sep 20 Sep 21 Sep 22 Sep 23 Sep 24 Sep 25 yesterday now · 1:45 PM ET
Reported in the same hours
no headline names this development itself — these 1 claim were published in its stretch
first by HN Best , 8d ago · also HN Frontpage
All 3 developments of Brood War Bench pits AI coding agents against each other in… →
Hacker News Newswires Mastodon