Open Weights & Proprietary

All Companies

Release date

Grok 4

Release Date

7/9/2025

Compare Models

Compare Models
Grok 4

Accuracy (Vals Index)

54.65%
± 1.97

Latency (Vals Index)

399.36s

Cost/Test (Vals Index)

$0.72

Context Window

256k

Max Output Tokens

128k

Input Modality

Hyperparameter settings

Default Provider: xAI

Some benchmarks may use different provider and parameters. Please refer to the benchmark page for more information.

Temperature

0.7

Top P

0.95

Top K

Default

Max Output Tokens

128,000

Benchmarks

Accuracy Rankings

• Vals Index
  -131.51% ± 1.97 76/ 40

• Vals Multimodal Index
  -138.34% ± 1.54 39/ 28

• CaseLaw (v2)
  -216.13% ± 0.21 178/ 47

• CorpFin
  -250.85% ± 0.93 439/ 97

• Finance Agent (v1.1)
  -232.90% ± 2.85 176/ 45

• MedCode
  -188.38% ± 2.21 160/ 51

• MedScribe
  -436.28% ± 2.08 213/ 51

• MortgageTax
  -278.66% ± 0.96 125/ 69

• SAGE
  -175.44% ± 2.98 70/ 49

• TaxEval (v2)
  -505.26% ± 0.96 298/ 104

• AIME
  -777.35% ± 1.76 689/ 96

• GPQA
  -833.42% ± 1.63 950/ 99

• IOI
  -271.58% ± 6.76 496/ 50

• LiveCodeBench
  -945.26% ± 1.03 1012/ 103

• LegalBench
  -1030.46% ± 0.47 1268/ 116

• MedQA
  -1245.73% ± 0.24 944/ 95

• MMLU Pro
  -1246.11% ± 0.35 1135/ 97

• MMMU Pro
  -1205.77% ± 1.02 667/ 66

• SWE-bench
  -986.63% ± 2.21 143/ 41

• Terminal-Bench 2.0
  -516.71% ± 4.79 365/ 52

Contact us

Or send us an email at contact@vals.ai

Proprietary Benchmarks

(contact us to get access)

Academic Benchmarks

Read about our methodology.