02 — Run · Round 2 · Round robin · r2-rr-sonnet5-gpt56luna-g3
Claude Sonnet 5 vs GPT-5.6 Luna
Claude Sonnet 5 destroyed the enemy base after 7.2 game-minutes · 10.1 min wall clock
WIN ukraine · west
Claude Sonnet 5
anthropic/claude-sonnet-5
Value destroyed$13,700Value lost$1,600
Units killed / lost19 / 16Buildings killed / lost7 / 0
Peak army$6,700Orders issued210
Decision turns (failed)55 (1)Mean latency4.5sModel cost$0.390
LOSS russia · east
GPT-5.6 Luna
openai/gpt-5.6-luna
Value destroyed$1,600Value lost$13,700
Units killed / lost16 / 19Buildings killed / lost0 / 7
Peak army$2,500Orders issued59
Decision turns (failed)55 (0)Mean latency3.6sModel cost$0.044
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 | sonnet5 won: base destroyed |
| pass | code_check | decisions-logged | turns per slot: {'Multi0': 55, 'Multi1': 55} |
| pass | code_check | llm-turns-valid | failed/total per slot: {'Multi0': '0/55', 'Multi1': '0/55'} |
| pass | code_check | video-decodes | duration 74.4s |
| pass | code_check | video-complete | video ends at tick 10860 of 10860 |
| pass | code_check | limits-respected | ended after 10.1 wall-minutes (budget 30.0 min): base destroyed |
| pass | checklist | notes-explain-win | Notes section explains Claude Sonnet 5 won by destroying base with superior army; cites both models final thoughts from decisions.jsonl |
| pass | checklist | players-match-request | result.json confirms Multi0=sonnet5 (anthropic/claude-sonnet-5) and Multi1=gpt56luna (openai/gpt-5.6-luna) matching the requested players |
| pass | checklist | no-silent-outage | all 55 decision ticks logged at regular 200-tick intervals with zero LLM failures for both players |
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.