Best AI models by Effective Value (π)
Ranked by Effective Value (π) β the single metric that fuses accuracy, speed, cost, and the exponential error-cascade of deep agentic work. π rewards models that complete long multi-step chains without human rescue. For production agents (not one-shot chat), π is the number that matters most. Shown indexed to the top model = 100 (raw π spans a ~2.7MΓ range).
Ranking 22 models across the meo-tested roster Β· as of 2026-06-08.
| # | Model | Lab | π index |
|---|---|---|---|
| 1 | Anthropic: Claude Opus 4.8 | anthropic | 100.0 |
| 2 | OpenAI: GPT-5.5 | openai | 73.6 |
| 3 | xAI: Grok 4.3 | x-ai | 21.2 |
| 4 | Google: Gemini 3.5 Flash | 16.3 | |
| 5 | Google: Gemini 3.1 Pro Preview | 7.8 | |
| 6 | inclusionAI: Ring-2.6-1T | inclusionai | 5.9 |
| 7 | NVIDIA: Nemotron 3 Ultra | nvidia | 0.73 |
| 8 | Xiaomi: MiMo-V2.5 | xiaomi | 0.68 |
| 9 | Qwen: Qwen3.7 Max | qwen | 0.58 |
| 10 | MoonshotAI: Kimi K2.6 | moonshotai | 0.32 |
| 11 | DeepSeek: DeepSeek V4 Flash | deepseek | 0.30 |
| 12 | Perceptron: Perceptron Mk1 | perceptron | 0.26 |
| 13 | MiniMax: MiniMax M3 | minimax | 0.22 |
| 14 | DeepSeek: DeepSeek V4 Pro | deepseek | 0.17 |
| 15 | Qwen: Qwen3.7 Plus | qwen | 0.10 |
| 16 | Xiaomi: MiMo-V2.5-Pro | xiaomi | 0.095 |
| 17 | Owl Alpha | openrouter | 0.071 |
| 18 | Google: Gemini 3.1 Flash Lite | 0.052 | |
| 19 | Z.ai: GLM 5.1 | z-ai | 0.003 |
| 20 | StepFun: Step 3.7 Flash | stepfun | <0.001 |
| 21 | Tencent: Hy3 preview | tencent | <0.001 |
| 22 | Arcee AI: Trinity Large Thinking | arcee-ai | <0.001 |