02 — Run · Round 2 · Round robin · r2-rr-gpt56terra-gpt56luna-g2
GPT-5.6 Luna vs GPT-5.6 Terra
GPT-5.6 Terra destroyed the enemy base after 5.9 game-minutes · 7.7 min wall clock
LOSS russia · west
GPT-5.6 Luna
openai/gpt-5.6-luna
Value destroyed$2,600Value lost$9,600
Units killed / lost20 / 15Buildings killed / lost0 / 6
Peak army$1,700Orders issued69
Decision turns (failed)45 (0)Mean latency3.2sModel cost$0.034
WIN russia · east
GPT-5.6 Terra
openai/gpt-5.6-terra
Value destroyed$9,600Value lost$2,600
Units killed / lost15 / 20Buildings killed / lost6 / 0
Peak army$4,000Orders issued84
Decision turns (failed)45 (0)Mean latency3.4sModel cost$0.322
Evaluation
passchecks 10/10 verdict computed from the gating checks
| Status | Kind | Check | Detail |
|---|---|---|---|
| pass | step | environment | Environment check (snapshot, engine, content, key) |
| pass | step | launch | Launch run-match as a detached process |
| pass | step | supervise | Supervise in the foreground until run-match exits |
| pass | step | render | Render the 1080p recording |
| pass | step | notes | Append Notes to summary.md |
| pass | code_check | outputs-exist | all present |
| pass | code_check | game-has-winner | gpt56terra won: base destroyed |
| pass | code_check | decisions-logged | turns per slot: {'Multi0': 45, 'Multi1': 45} |
| pass | code_check | llm-turns-valid | failed/total per slot: {'Multi0': '0/45', 'Multi1': '0/45'} |
| pass | code_check | video-decodes | duration 60.7s |
| pass | code_check | video-complete | video ends at tick 8801 of 8801 |
| pass | code_check | limits-respected | ended after 7.7 wall-minutes (budget 30.0 min): base destroyed |
| pass | checklist | notes-explain-win | summary.md Notes section cites final thoughts from both models and explains Terra won via economic dominance and base destruction |
| pass | checklist | players-match-request | result.json records gpt56luna as Multi0 and gpt56terra as Multi1 exactly as requested |
| pass | checklist | no-silent-outage | run.log shows uninterrupted decision turns with 0 LLM failures across all 45 turns per player |
Decision timeline
What each model saw as its situation and what it ordered, every 8 game-seconds. Click a turn to jump the video there.