TIS-100 Bench: Signal Pattern Detector

40196 Easy Streaming state Built-in puzzle · TIS-100 Bench report →

Detect three consecutive zeros in a stream. Overlapping matches make the update and reset rules matter.

In TIS-100 Bench, 16 LLMs each had one attempt at this TIS-100 puzzle, with 10 minutes to find a working program and keep improving it, without seeing the community leaderboard. 15 of 16 models solved it. No model found a new Pareto-frontier point beyond the community leaderboard.

15 / 16models solved it
126best LLM cycles (82 community)
0new frontier points

Records

ScoreCommunity bestBest LLMModel
Cycles82126GPT-6.1 Sol · +54%
Nodes44DeepSeek V4.1 Flash Ties record
Instructions1313GPT-6.1 Sol Ties record

Community records from the community leaderboard when we froze it. A record counts only after fresh-game replay and 100,000 passing random tests.

Pareto frontier

Pareto explorer

Lower is better ↙
Loading interactive Pareto explorer…
Community leaderboardModel frontier pointNew frontier point
Select a point with the mouse or keyboard to inspect its scores and program.

Search over time

Loading puzzle trajectories…

Every model

ModelSolvedBest cyclesBest nodesBest instructionsNew frontier points
01GPT-6.1 SolYes126413—
02Kimi K3.1 ProYes157414—
03Sonnet 5.5Yes159415—
04GPT-6 AstraYes160414—
05Gemini 3.8 FlashYes160415—
06Opus 5.5Yes160413—
07Gemini 4 ArgonYes162413—
08Haiku 5.5Yes167414—
09Muse Spark 1.3Yes167414—
10Grok 4.7Yes170417—
11Fable 5.1Yes191416—
12GLM-5.3Yes200417—
13Muse Wonder 1.0Yes226416—
14DeepSeek V4.1 FlashYes232415—
15GPT-6 LunaYes282417—
16MiMo-V2.6-ProNo————

Best scores can come from different programs. New frontier points are not matched or beaten on all three scores by any community program.

Best model programs

These programs are in TIS-100 save format and released under the MIT License. All of this puzzle’s model programs are on GitHub and Hugging Face.

GPT-6.1 Sol Fewest cycles

126 cycles · 5 nodes · 28 instructions40196.126-5-28.txt on GitHub

GPT-6.1 Sol program running on 40196 in TIS-100
Node 1
MOV UP RIGHT
MOV UP DOWN
Node 2
A: MOV LEFT ACC
JEZ Z
MOV 999 DOWN
JMP A
Z: MOV 1 DOWN
Node 4
A: MOV UP ACC
JEZ Z
MOV 999 RIGHT
JMP A
Z: MOV 1 RIGHT
Node 5
MOV UP DOWN
MOV LEFT DOWN
Node 9
JRO UP
MOV 0 DOWN
JRO UP
MOV 0 DOWN
C: JRO UP
MOV 1 DOWN
JRO UP
MOV 1 DOWN
JRO UP
MOV 1 DOWN
JRO UP
MOV 1 DOWN
JMP C
MOV 0 DOWN

Kimi K3.1 Pro Fewest nodes

157 cycles · 4 nodes · 15 instructions40196.157-4-15.txt on GitHub

Kimi K3.1 Pro program running on 40196 in TIS-100
Node 1
S0: MOV 0, RIGHT
MOV UP, ACC
JNZ S0
S1: MOV 0, RIGHT
MOV UP, ACC
JNZ S0
E: MOV UP, ACC
JNZ Z
O: MOV 1, RIGHT
R: MOV UP, ACC
JEZ O
Z: MOV 0, RIGHT
Node 2
MOV LEFT, DOWN
Node 5
MOV UP, DOWN
Node 9
MOV UP, DOWN

GPT-6.1 Sol Fewest instructions

179 cycles · 4 nodes · 13 instructions40196.179-4-13.txt on GitHub

GPT-6.1 Sol program running on 40196 in TIS-100
Node 1
A: MOV UP ACC
O: MOV 0 DOWN
JNZ A
MOV UP ACC
MOV 0 DOWN
JNZ A
C: MOV UP ACC
JNZ O
MOV 1 DOWN
JMP C
Node 4
MOV UP RIGHT
Node 5
MOV LEFT DOWN
Node 9
MOV UP DOWN