complex_qwen3.6-35b_exact68_final2
Qwen3.6-35B on the complex floor, 5 episodes. Cited by the thesis in:
| Wins (boss defeated, alive) | 2/5 [0.12, 0.77] |
| Final HP / exact optimum, mean over wins | 85% |
| Died | 1 |
| Final HP, mean | 31 |
| Seconds per model call | 32.2 |
| Parse error rate | 4.5% |
| Floor | Complex floor |
| Model | Qwen3.6-35B (qwen3.6-35b-a3b, thinking) |
| Planning | reactive |
| Prompt | complex_exact68_sequence.txt |
| Temperature | 0 |
| Lookahead depth | 3 |
| Mechanical overrides | off (all engine replacements and assists disabled) |
Engine settings (environment, 5)
MOTA_SUPPRESS_HINTS=0 THINKING_MAX_TOKENS=16384 CONVERSATION_HISTORY_TURNS=0 MOTA_ORDER_TWO_SIDED=1 MOTA_FIGHT_KEY_FLOOD=1
In a checkout of mt-ai-lab with an API key set (the result goes to a new file rerun_<time>_complex_qwen3.6-35b_exact68_final2, never over the thesis result):
python scripts/run_catalog.py --codes X68S
Result file: results/reliability/complex_qwen3.6-35b_exact68_final2.json · updated 2026-09-29