Introducing Parallel Search Fast: A web search API for agents that’s fast, cheap, and accurate. Learn more.[Learn more.]
Search Capability Leaderboard

Gemini 3.7 Flash

GoogleClosed

Evaluated

#17 in Search Intelligence · #5 in Search Efficiency

Search Intelligence Score

66.8

+27.4 lift from search

050100
Lift from search
+27.4
39.4 without search
Cost per 1K tasks
$130
#5 in Search Efficiency

Accuracy by benchmark

Accuracy on each eval, with search (solid) and without (light).

  • DSQA87.5 vs 58.3
  • HLE46.0 vs 42.0
  • WISER67.0 vs 18.0

Cost per 1K tasks

How the cost splits between model inference, search calls, and page extraction.

Inference
$117
90%
Search
$7.61
6%
Extract
$4.92
4%

Compared with

Other models from the same provider alongside models that score closest.

Search Intelligence Score averages accuracy on DSQA, HLE, and WISER with equal weight; cost is averaged the same way, with cost shown per 1,000 tasks. Evaluated . Leaderboard last updated . Full methodology.

# Gemini 3.7 Flash · Search Capability Leaderboard

  • Search Intelligence Score: 66.8 (#17 in Search Intelligence)
  • Without search: 39.4 · Lift from search +27.4
  • Cost per 1K tasks: $130 (#5 in Search Efficiency)
  • Evaluated:
  • Lab: Google · Weights: Closed · Model id: google/gemini-3.7-flash

## Accuracy by benchmark

SuiteWith searchWithout search
DSQA87.558.3
HLE46.042.0
WISER67.018.0

## Compared with

ModelScoreLiftCost per 1K tasks
Gemini 3.7 Flash66.8+27.4$130
Gemini 3.8 Flash65.1+26.1$177
Gemini 4 Argon72.2+37.2$270
Gemini 3 Flash58.6+24.5$114
GPT-6 Sol66.6+28.7$183

Leaderboard last updated . Methodology: /leaderboard#methodology