benchmarks.site

Public leaderboard · Aggregated leaderboards

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.

Leaderboard

Top 3 right now

  1. 1stClaude Fable 5.271.31
  2. 2ndClaude Sonnet 5.570.85
  3. 3rdGemini 4 Carbon70.45

What it measures

Independent evaluations on industry tasks such as legal, finance, tax, medical and coding work, many with private test sets.

How to read the numbers

Vals Index accuracy: average performance on finance, coding, legal and tax tasks, weighted by sector (%) — higher is better.

Results

25 rows · higher is better (Accuracy (%))CSV
1Claude Fable 5.271.31$17.9443m42s
2Claude Sonnet 5.570.85$16.181h26m
3Gemini 4 Carbon70.45$10.5147m45s
4Claude Opus 5.570.14$24.771h25m
5Gemini 4 Argon69.39$13.3744m58s
6Gemini 4 Hydrogen68.49$9.8432m05s
7Wonder Full66.79$6.8159m19s
8GPT-6 Astra65.02$21.2833m39s
9Wonder64.88$10.6535m24s
10GPT-6.1 Sol63.94$3.6144m24s
11Qwen 4 Max63.20$9.331h05m
12Kimi K3.1 Pro62.52$6.441h04m
13DeepSeek V4.1 Flash 101761.26$0.6639m55s
14GPT-6.1 Terra60.54$2.9439m43s
15Muse Ember60.48$0.1515m02s
16GLM 5.359.85$8.521h04m
17Hy459.35$1.7251m31s
18Qwen 3.8 Max 090259.05$8.731h46m
19Claude Haiku 5.558.73$3.1749m57s
20Grok 4.758.62$11.6333m34s
21MiMo V2.6 Pro57.81$0.431h13m
22Mistral Large 456.53$3.7247m48s
23Inkling 2 Small56.08$3.5150m35s
24Gemini 3.8 Flash55.95$5.0335m58s
25MiMo V2.6 Flash54.83$0.221h07m

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

Source: Vals AI · vals.ai · results as of 2026-11-15 · retrieved 2026-11-17

Markdown · More Aggregated leaderboards