LLM benchmark scores against price
The top 20 models on the Artificial Analysis Intelligence Index, plotted against what a million tokens costs. Free Opencode Zen models get their own lane.
Artificial Analysis Index v5 · updated about 2 hours ago
Latest frontier lab releases
12 Nov 2026Index v5 · fetched 18:45 UTC
33 / 3329 Oct 202612 Nov 2026
- Entered the rankingGemini 4 Hydrogen
- Dropped outDeepSeek V4.1 Flash Free, DeepSeek V4.1 Flash 1017
- Anthropic
- OpenAI
- Other labs
- Free on Opencode Zen
- Pareto frontier
CursorBench 4.0
Real coding tasks run in the Cursor agent. Each line is one model across its reasoning efforts. Higher and further right is better. Pick models below the chart. Source
All models
21 models · 12 Nov 2026| # | Model | Intelligence | € / 1M | Move | Hide | |
|---|---|---|---|---|---|---|
| 1 | Anthropic | 64.6 | 22.08 | – | ||
| 2 | 60.0 | 11.04 | – | |||
| 3 | Anthropic | 59.1 | 11.04 | – | ||
| 4 | GPT-6 AstraMax | OpenAI | 56.9 | 27.60 | – | |
| 5 | OpenAI | 54.6 | 5.52 | – | ||
| 6 | Anthropic | 54.5 | 5.52 | – | ||
| 7 | 53.9 | 4.14 | new | |||
| 8 | OpenAI | 51.3 | 2.76 | 1 | ||
| 9 | Gemini 4 ArgonHigh | 50.4 | 11.04 | 1 | ||
| 10 | Claude Fable 5.1Xhigh | Anthropic | 50.0 | 27.60 | 1 | |
| 11 | Kimi | 46.5 | 5.52 | 1 | ||
| 12 | GPT-6 SolMax | OpenAI | 44.2 | 5.52 | 1 | |
| 13 | Anthropic | 44.2 | 13.80 | 1 | ||
| 14 | Muse WonderMax | Meta | 43.6 | 11.04 | 1 | |
| 15 | OpenCode Zen | 36.6 | free | 1 | ||
| 16 | Meta | 36.6 | 2.53 | 1 | ||
| 17 | Anthropic | 36.5 | 0.28 | 1 | ||
| 18 | Meta | 36.0 | 0.28 | 1 | ||
| 19 | Grok 4.7Xhigh | SpaceXAI | 35.5 | 3.68 | 1 | |
| 20 | Tencent Hy4new | Tencent | 34.6 | 1.53 | 1 | |
| 21 | Qwen3.8 Max0902 | Alibaba | 33.7 | 3.68 | 1 |