Benchmarks · agentic
Web Search Index
Compares native model-provider search with independent web search tools on finance and legal research tasks.
4 models · 4 results · updated 2026-10-21 · vals.ai ↗
Leaderboard
Fable 5 max
More from this benchmark
1 published measuresTime per task9 results · lower is better
All results
4 rows| # | Model | Effort | Web Search Index | Subscription $ / task | API $ / task | Time | Harness |
|---|---|---|---|---|---|---|---|
| 1 | Fable 5 | max | 46.9% | $0.49 | $4.37 | 12 min | Native |
| 2 | GPT-5.6 Sol | max | 43.6% | $0.053 | $0.95 | 5.5 min | Native |
| 3 | Gemini 3.5 Flash | high | 37.0% | $0.50 | $0.50 | 4.2 min | Native |
| 4 | Grok 4.5 | high | 35.5% | $0.16 | $0.99 | 2.3 min | Native |