TIS-100 Bench: T20 Node Emulator

NEXUS.22.280.8 Medium Memory & interpreters Streaming state Built-in puzzle · TIS-100 Bench report →

Interpret commands for a tiny machine with two stored values: load, swap, add, and output. Command dispatch must preserve machine state.

In TIS-100 Bench, 16 LLMs each had one attempt at this TIS-100 puzzle, with 30 minutes to find a working program and keep improving it, without seeing the community leaderboard. 12 of 16 models solved it. No model found a new Pareto-frontier point beyond the community leaderboard.

12 / 16models solved it
182best LLM cycles (168 community)
0new frontier points

Records

ScoreCommunity bestBest LLMModel
Cycles168182Gemini 4 Argon · +8%
Nodes44Fable 5.1 Ties record
Instructions2222Fable 5.1 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
01Gemini 4 ArgonYes182422—
02GPT-6.1 SolYes184422—
03GPT-6 AstraYes191422—
04Fable 5.1Yes192422—
05Opus 5.5Yes211422—
06Haiku 5.5Yes213623—
07Grok 4.7Yes214428—
08Kimi K3.1 ProYes257426—
09Sonnet 5.5Yes262433—
10Muse Wonder 1.0Yes279533—
11Gemini 3.8 FlashYes314432—
12DeepSeek V4.1 FlashYes510654—
13GLM-5.3No————
14MiMo-V2.6-ProNo————
15Muse Spark 1.3No————
16GPT-6 LunaNo————

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.

Gemini 4 Argon Fewest cycles

182 cycles · 9 nodes · 26 instructionsNEXUS.22.280.8.182-9-26.txt on GitHub

Gemini 4 Argon program running on NEXUS.22.280.8 in TIS-100
Node 0
MOV RIGHT, ACC
SUB 5
MOV ACC, DOWN
Node 1
MOV UP, ACC
ADD ACC
MOV ACC, LEFT
Node 2
MOV UP, DOWN
Node 4
MOV UP, RIGHT
Node 5
MOV LEFT, RIGHT
Node 6
JMP L
MOV UP, ACC
JMP L
SWP
MOV UP, ACC
SWP
L: JRO LEFT
SWP
JMP P3
MOV ACC, DOWN
JMP L
P3: MOV ACC, RIGHT
SWP
ADD RIGHT
Node 7
MOV LEFT, LEFT
Node 9
MOV RIGHT, DOWN
Node 10
MOV UP, LEFT

Gemini 4 Argon Fewest nodes

300 cycles · 4 nodes · 25 instructionsNEXUS.22.280.8.300-4-25.txt on GitHub

Gemini 4 Argon program running on NEXUS.22.280.8 in TIS-100
Node 1
L: MOV UP, ACC
ADD ACC
SUB 5
MOV ACC, RIGHT
JLZ L
JRO ACC
MOV RIGHT, RIGHT
JMP L
MOV RIGHT, DOWN
Node 2
JMP L
MOV UP, ACC
JMP L
SWP
MOV UP, ACC
SWP
L: JRO LEFT
SWP
JMP P3
MOV ACC, LEFT
JMP L
P3: MOV ACC, LEFT
SWP
ADD LEFT
Node 5
MOV UP, DOWN
Node 9
MOV UP, DOWN

Fable 5.1 Fewest instructions

211 cycles · 6 nodes · 22 instructionsNEXUS.22.280.8.211-6-22.txt on GitHub

Fable 5.1 program running on NEXUS.22.280.8 in TIS-100
Node 1
MOV UP, ACC
ADD ACC
ADD 1
MOV ACC, DOWN
Node 2
MOV UP, DOWN
Node 4
MOV RIGHT, RIGHT
Node 5
J: JRO UP
MOV RIGHT, ACC
JMP J
SWP
MOV RIGHT, ACC
SWP
JMP J
SWP
JMP A
MOV ACC, DOWN
JMP J
A: MOV ACC, LEFT
SWP
ADD LEFT
Node 6
MOV UP, LEFT
Node 9
MOV UP, DOWN