# Vals AI

Vals AI tests AI models on realistic professional work in finance, law, tax and coding and combines the results into one index, aiming to show how useful models are for real jobs.

Category: aggregators. Type: leaderboard. Status: active.
Metric: Vals Index accuracy: average performance on finance, coding, legal and tax tasks, weighted by sector (%) (higher is better).
Source: Vals AI — https://www.vals.ai/benchmarks/vals_index, results as of 2026-11-15, retrieved 2026-11-17, license: not stated.

| Rank | Model | Accuracy (%) | Cost / test | Duration / test |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 | 71.31 | $17.94 | 43m42s |
| 2 | Claude Sonnet 5.5 | 70.85 | $16.18 | 1h26m |
| 3 | Gemini 4 Carbon | 70.45 | $10.51 | 47m45s |
| 4 | Claude Opus 5.5 | 70.14 | $24.77 | 1h25m |
| 5 | Gemini 4 Argon | 69.39 | $13.37 | 44m58s |
| 6 | Gemini 4 Hydrogen | 68.49 | $9.84 | 32m05s |
| 7 | Wonder Full | 66.79 | $6.81 | 59m19s |
| 8 | GPT-6 Astra | 65.02 | $21.28 | 33m39s |
| 9 | Wonder | 64.88 | $10.65 | 35m24s |
| 10 | GPT-6.1 Sol | 63.94 | $3.61 | 44m24s |
| 11 | Qwen 4 Max | 63.20 | $9.33 | 1h05m |
| 12 | Kimi K3.1 Pro | 62.52 | $6.44 | 1h04m |
| 13 | DeepSeek V4.1 Flash 1017 | 61.26 | $0.66 | 39m55s |
| 14 | GPT-6.1 Terra | 60.54 | $2.94 | 39m43s |
| 15 | Muse Ember | 60.48 | $0.15 | 15m02s |
| 16 | GLM 5.3 | 59.85 | $8.52 | 1h04m |
| 17 | Hy4 | 59.35 | $1.72 | 51m31s |
| 18 | Qwen 3.8 Max 0902 | 59.05 | $8.73 | 1h46m |
| 19 | Claude Haiku 5.5 | 58.73 | $3.17 | 49m57s |
| 20 | Grok 4.7 | 58.62 | $11.63 | 33m34s |
| 21 | MiMo V2.6 Pro | 57.81 | $0.43 | 1h13m |
| 22 | Mistral Large 4 | 56.53 | $3.72 | 47m48s |
| 23 | Inkling 2 Small | 56.08 | $3.51 | 50m35s |
| 24 | Gemini 3.8 Flash | 55.95 | $5.03 | 35m58s |
| 25 | MiMo V2.6 Flash | 54.83 | $0.22 | 1h07m |

Notes: Vals Index v2.2 only; Vals hosts many other benchmarks. Read from the rendered page table.
