Best AI models for tape-machine simulation
Ranked by accuracy on the meo “tape-machine simulation” domain — one of eleven un-leaked, contamination-resistant domains. Per-domain strength reveals where a model genuinely reasons versus where it pattern-matches; the leaders differ sharply by domain.
Ranking 22 models across the meo-tested roster · as of 2026-06-08.
| # | Model | Lab | accuracy |
|---|---|---|---|
| 1 | OpenAI: GPT-5.5 | openai | 100% |
| 2 | Anthropic: Claude Opus 4.8 | anthropic | 100% |
| 3 | Google: Gemini 3.1 Pro Preview | 100% | |
| 4 | Google: Gemini 3.5 Flash | 100% | |
| 5 | Qwen: Qwen3.7 Max | qwen | 100% |
| 6 | xAI: Grok 4.3 | x-ai | 100% |
| 7 | DeepSeek: DeepSeek V4 Flash | deepseek | 100% |
| 8 | MoonshotAI: Kimi K2.6 | moonshotai | 100% |
| 9 | NVIDIA: Nemotron 3 Ultra | nvidia | 100% |
| 10 | Xiaomi: MiMo-V2.5-Pro | xiaomi | 100% |
| 11 | Perceptron: Perceptron Mk1 | perceptron | 100% |
| 12 | inclusionAI: Ring-2.6-1T | inclusionai | 93% |
| 13 | DeepSeek: DeepSeek V4 Pro | deepseek | 93% |
| 14 | Qwen: Qwen3.7 Plus | qwen | 93% |
| 15 | MiniMax: MiniMax M3 | minimax | 93% |
| 16 | Xiaomi: MiMo-V2.5 | xiaomi | 87% |
| 17 | Owl Alpha | openrouter | 87% |
| 18 | Z.ai: GLM 5.1 | z-ai | 53% |
| 19 | Tencent: Hy3 preview | tencent | 33% |
| 20 | Google: Gemini 3.1 Flash Lite | 27% | |
| 21 | StepFun: Step 3.7 Flash | stepfun | 0% |
| 22 | Arcee AI: Trinity Large Thinking | arcee-ai | 0% |