Qwen 3.6 27B NVFP4 Instruct | Benchmark #189

Described in:Bomberman-like game built by Qwen 3.6 and Qwen 3.8 as agents

Benchmark summary

Bomberman-like game built in SC# as an agent

ModelQwen 3.6 27B QuantizationNVFP4 ModeInstruct

The model builds a Bomberman-like maze bomber game in SC# as an agent: it writes the app, compiles it, runs its own tests and fixes the code. Hidden tests check the rules from the task; each hidden test is its own result.

Score 15/15 100.0% passed
15 passed15 total 1 AI query 50 tool calls TOKENS IN3 408 498 TOKENS OUT36 128

Results by difficulty

1 group
Level Unknown / other
15 / 15 100.0%
passed
1 AI query 50 tool calls IN3 408 498 OUT36 128

Results by category

1 group
Category Other
100.0% 15/15 passed
1 AI query 50 tool calls TOKENS IN3 408 498 TOKENS OUT36 128

Tests with multiple category tags are counted in each matching category, so category totals may exceed the report total.

Benchmark items
Script / test Status Tokens
Building Games in SC#: Bomberman-like maze bomber 15 passed
100.0%
IN3 408 498 OUT36 128
✓
BombermanThe model builds a Bomberman game from the task as an agent; this item passes when the app compiles and can be exported, and each hidden test is its own item.1 AI query 50 tool calls
Passed IN3 408 498 OUT36 128
✓
Hidden_ChainReactionA bomb in the fire explodes in the same frame, and the fire stops at it.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_EnemiesAn enemy in the fire dies and gives 100 points; an enemy on the player's cell costs the player a life.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_EnemiesWalkEnemies take a step every 12 frames: each moves to a neighbor cell that is free for walking, never into walls, bricks or bombs.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_ExplosionCrossA bomb explodes 90 frames after it is placed; the fire is a cross of bombRange cells in every direction.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_FireLasts15FramesFire lasts 15 frames: it is gone 15 frames after the explosion.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_InitialStateAfter AppInit, the walls, the bricks, the player, the enemies and the counters have their starting values.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_PlacingBombsSpace places a bomb on the player's cell; the player can walk off it but not back onto it, and only maxBombs bombs can be placed.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_PlayerInTheFireThe player in the fire loses a life and goes back to (1, 1); with the last life the game is over.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_PlayerMovesThe player moves one cell when an arrow key is held, then waits 6 frames before the next move, and walls stop it.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_PowerUpsPicking up a power-up adds a bomb or lengthens the fire.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_ScreenShowsFireAndEnemiesFire and enemies are drawn on their cells.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_ScreenShowsTheSceneThe screen shows the board area, walls, the player, a bomb and the texts.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_WallsAndBricksStopFireFire stops before walls, destroys the first brick in its way and stops there.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0
✓
Hidden_WinningAndRestartWhen the last enemy dies the game is won and stops; Enter starts a new game.0 AI queries 0 tool callsHiddenTest
Passed IN0 OUT0