Benchmarks / External Composite Indices
Vals Index
Vals AI's GDP-weighted composite over professional-domain benchmarks. Displayed for reference only.
maintained by Vals AI · 35 observations on record
Versions
1 version
| Version | Tasks |
|---|---|
| Vals Index current vals-index | — |
Current leaderboard
Vals Index · Index points · top 25 of 35 systems
| # | System | Index points |
|---|---|---|
| 1 | Claude Fable 5max effort Anthropic | 75.1 points |
| 2 | GPT 5.6 Solmax effort OpenAI | 72.6 points |
| 3 | Claude Opus 4.8max effort Anthropic | 70.4 points |
| 4 | Claude Sonnet 5max effort Anthropic | 68.6 points |
| 5 | Muse Spark 1.1xhigh effort Meta AI | 68.4 points |
| 6 | GPT 5.5 (2026-04-22)xhigh effort OpenAI | 68.0 points |
| 7 | Claude Opus 4.7high effort Anthropic | 66.1 points |
| 8 | Grok 4.5high effort xAI | 65.3 points |
| 9 | GPT 5.6 Terraxhigh effort OpenAI | 65.1 points |
| 10 | GLM 5.2max effort Z.ai | 65.0 points |
| 11 | Gemini 3.5 Flashhigh effort Google DeepMind | 62.7 points |
| 12 | Claude Sonnet 4.6max effort Anthropic | 60.1 points |
| 13 | Minimax M3 MiniMax | 58.9 points |
| 14 | Qwen3 7 Max Alibaba (Qwen) | 57.5 points |
| 15 | Deepseek V4 Promax effort DeepSeek | 55.6 points |
| 16 | Kimi K2 6 Moonshot AI | 55.2 points |
| 17 | Gemini 3.1 Pro Previewhigh effort Google DeepMind | 53.8 points |
| 18 | GLM 5.1 Z.ai | 52.5 points |
| 19 | GPT 5.4 Mini (2026-03-17)xhigh effort OpenAI | 52.4 points |
| 20 | Qwen3 7 Plus Alibaba | 52.3 points |
| 21 | Mimo V2 5 xiaomi | 51.6 points |
| 22 | Mimo V2 5 Pro xiaomi | 50.7 points |
| 23 | Gemini 3 Flash Previewhigh effort Google DeepMind | 49.5 points |
| 24 | Qwen3 6 Plus Alibaba | 48.9 points |
| 25 | GPT 5.4 Nano (2026-03-17)high effort OpenAI | 46.6 points |
One row per evaluated system — reasoning-effort variants rank separately.