Inference
EN
Vals AI

Vals Web Search Index

Comparing native provider search against independent web-search tools on legal-research and finance-analysis tasks

ModelsModels with a result.12
Top resultBest result on this benchmark.57.5%Claude Opus 5.5 Parallel Fast
Top-3 spreadPoints from first to third.3.8 pts
UpdatedDate the source changed the results.4 Nov 2026

Result and cost

12 of 12 results

Results

12 results
#ModelResultThe score from the source.CostUS dollars to run one task.TimeTime to run one task.
01Claude Opus 5.5 Parallel FastParallel Fast57.5% ±2.0$2.277 min
02Claude Opus 5.5 Parallel AdvancedParallel Advanced54.1% ±2.0$2.217 min
03Claude Opus 5.5Native53.7% ±2.0$3.4810 min
04GPT-5.6 Sol KeenableKeenable49.1% ±2.0$8.847 min
05Claude Fable 5 ExaExa48.5% ±2.0$3.486 min
06Claude Fable 5Native46.9% ±2.1$4.3712 min
07GPT-5.6 Sol ExaExa45.2% ±2.0$18.979 min
08GPT-5.6 SolNative43.6% ±2.0$0.9545 min
09Gemini 3.5 Flash ExaExa41.4% ±1.9$1.183 min
10Grok 4.5 ExaExa38.7% ±1.9$0.5762 min
11Gemini 3.5 FlashNative37.0% ±1.9$0.4994 min
12Grok 4.5Native35.5% ±1.9$0.9912 min