Models30 scored

Shopping scores

How each model performs when the Agent drives it through a funnel on a real store. Every rate below comes from completed sessions — a tool call, a cart, a checkout attempt — never an estimate.

Compare
GoogleGoogleGemini 3.8 Flash
52Fair
Checkout6.7
Cart53.3
Search100
xAIxAIGrok 4.3 retired
43Fair
Checkout26.3
Cart33.1
Search64.4
DeepSeekDeepSeekDeepSeek V3.2
43Fair
Checkout20.7
Cart36.7
Search74.7
GoogleGoogleGemini 3.5 Flash
42Fair
Checkout38.1
Cart56
Search80.1
GoogleGoogleGemini 3.6 Flash
40Fair
Checkout1.1
Cart89.7
Search98.2
AlibabaAlibabaQwen 3.8 Flash
39Fair
Checkout22.2
Cart66.7
Search100
Most Reliable
xAIxAIGrok 4.5
39Fair
Checkout8.1
Cart37.8
Search86.5
AnthropicAnthropicClaude Sonnet 4.6
38Fair
Checkout34.7
Cart47.5
Search79.6
MetaMetaLlama 3.3 70B retired
37Fair
Checkout39.3
Cart52.7
Search82.1
Highest Checkout
GoogleGoogleGemini 2.5 Flash
36Fair
Checkout23.3
Cart38.5
Search74.9
DeepSeekDeepSeekDeepSeek V4.1 Flash
36Fair
Checkout14.3
Cart42.9
Search100
GoogleGoogleGemini 2.5 Pro
34Weak
Checkout19.8
Cart39.2
Search68.8
OpenAIOpenAIGPT-4o
32Weak
Checkout15
Cart36.7
Search72.5
MetaMetaMuse Spark 1.3
32Weak
Checkout0
Cart33.3
Search100
GoogleGoogleGemini 3.1 Pro
30Weak
Checkout19.5
Cart38.1
Search74.7
AnthropicAnthropicClaude Opus 4.8
29Weak
Checkout22.3
Cart31.6
Search63.6
GoogleGoogleGemini 3.7 Flash
29Weak
Checkout0
Cart21.4
Search92.9
OpenAIOpenAIGPT-6 Luna
27Weak
Checkout0
Cart7.7
Search100
AnthropicAnthropicClaude Sonnet 5
23Weak
Checkout7.7
Cart52.9
Search91.6
DeepSeekDeepSeekDeepSeek R1 (Reasoning)
23Weak
Checkout3.6
Cart8.3
Search38.1
xAIxAIGrok 3 Mini (Reasoning) retired
22Weak
Checkout11.1
Cart13.3
Search48.9
OpenAIOpenAIGPT-5.5
21Weak
Checkout10.2
Cart24.4
Search69.8
OpenAIOpenAIo4-mini (Reasoning)
21Weak
Checkout9.1
Cart28.9
Search78.5
DeepSeekDeepSeekDeepSeek V4 Flash
20Weak
Checkout7.5
Cart35
Search85
MetaMetaLlama 4 Maverick
20Weak
Checkout2.2
Cart6.7
Search35.6
Fastest
AlibabaAlibabaQwQ 32B (Reasoning) retired
19Weak
Checkout0
Cart7.1
Search28.6
Most Efficient
DeepSeekDeepSeekDeepSeek V4 Pro
18Weak
Checkout7.7
Cart35.9
Search76.9
AnthropicAnthropicClaude Opus 5
16Weak
Checkout8.9
Cart28.7
Search69.3
Moonshot AIMoonshot AIKimi K3
16Weak
Checkout3.1
Cart34.4
Search75
AlibabaAlibabaQwen 3.7 Max
16Weak
Checkout2.6
Cart35.9
Search76.9
Model choice matters. Store implementation matters more — run the same model against your own store to see which one you're looking at. Run Agent session