TIS-100 Bench: Self-Test Diagnostic
00150 Easy Routing & synchronization Built-in puzzle · 
TIS-100 Bench report →
Copy two input streams to their matching outputs. The first challenge is simply routing values through a damaged processor grid.
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. 16 of 16 models solved it. No model found a new Pareto-frontier point beyond the community leaderboard.
Records
| Score | Community best | Best LLM | Model |
|---|---|---|---|
| Cycles | 83 | 83 | DeepSeek V4.1 Flash Ties record |
| Nodes | 8 | 8 | DeepSeek V4.1 Flash Ties record |
| Instructions | 8 | 8 | DeepSeek V4.1 Flash 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 ↙Search over time
Loading puzzle trajectories…
The search over time
Best in-game score so far, before independent testing. Each score can come from a different program, and lines stop when the episode ends.
Every model
| Model | Solved | Best cycles | Best nodes | Best instructions | New frontier points |
|---|---|---|---|---|---|
| 01DeepSeek V4.1 Flash | Yes | 83 | 8 | 8 | — |
| 02Fable 5.1 | Yes | 83 | 8 | 8 | — |
| 03GLM-5.3 | Yes | 83 | 8 | 8 | — |
| 04GPT-6 Astra | Yes | 83 | 8 | 8 | — |
| 05GPT-6.1 Sol | Yes | 83 | 8 | 8 | — |
| 06Gemini 3.8 Flash | Yes | 83 | 8 | 8 | — |
| 07Gemini 4 Argon | Yes | 83 | 8 | 8 | — |
| 08Grok 4.7 | Yes | 83 | 8 | 8 | — |
| 09Haiku 5.5 | Yes | 83 | 8 | 8 | — |
| 10Kimi K3.1 Pro | Yes | 83 | 8 | 8 | — |
| 11MiMo-V2.6-Pro | Yes | 83 | 8 | 8 | — |
| 12Muse Spark 1.3 | Yes | 83 | 8 | 8 | — |
| 13Muse Wonder 1.0 | Yes | 83 | 8 | 8 | — |
| 14Opus 5.5 | Yes | 83 | 8 | 8 | — |
| 15Sonnet 5.5 | Yes | 83 | 8 | 8 | — |
| 16GPT-6 Luna | Yes | 83 | 8 | 8 | — |
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.
DeepSeek V4.1 Flash
83 cycles · 8 nodes · 8 instructions00150.83-8-8.txt on GitHub
MOV UP, DOWNMOV RIGHT, DOWNMOV UP, LEFTMOV UP, DOWNMOV UP, DOWNMOV UP, DOWNMOV UP, RIGHTMOV LEFT, DOWN