Sample dataThis tutorial uses sample data. Changes are not saved.

Compare

Variant A

gpt-6.1-sol

Variant B

claude-sonnet-5-5

Variant C

gemini-3.8-flash

Origin vs Sonnet 5.5 vs Gemini 3.8 Resolution rate

Origin: 64.00%Sonnet 5.5: 70.80%Gemini 3.8: 58.50%

Variant AOrigin· gpt-6.1-sol

Cost / request

$0.008930

Avg latency

1.04 s

Error rate

0.88%

Resolution rate

64.00%

Variant BSonnet 5.5· claude-sonnet-5-5

Cost / request

$0.005415

Avg latency

638 ms

Error rate

1.94%

Resolution rate

70.80%

Variant CGemini 3.8· gemini-3.8-flash

Cost / request

$0.007174

Avg latency

843 ms

Error rate

1.32%

Resolution rate

58.50%

Per-run comparison

One agent or workflow run

MetricOriginSonnet 5.5Gemini 3.8

Runs4,6888201,262

Requests / run2.14.02.6

Cost / run$0.018586$0.021244$0.018407

Node details

NodeOriginSonnet 5.5Gemini 3.8

search.answer$0.008930$0.005415$0.007174

Success-metric comparison

Resolution rate Resolved + helpful (70/30)Cost per resolution

Edit success metrics

EventWeightOriginSonnet 5.5Gemini 3.8

Resolved100%64.00%70.80%58.50%

CompositeΣ64.00%70.80%58.50%