UCP Playgroundo4-mini (Reasoning)/Models/o4-mini (Reasoning)
InspectorAgentDocs
OpenAIOpenAI

o4-mini (Reasoning)

OpenAI's compact reasoning model. Currently non-functional for shopping — 92.9% tool error rate with almost every tool call failing. Appears to struggle with MCP tool schemas entirely. Not recommended until OpenAI addresses the tool-use reliability issues.

Avg Tokens70,680Avg Duration39.8sTurns to Checkout11.5
CheckoutCartSearchTurnsTokens9.329.778617
Shopping Score21/100
Weak

Scores computed from real agent sessions against live UCP-enabled stores. Not estimated — every data point is from an actual tool call, checkout attempt, and store response.

Shopping Score Breakdown

Checkout Rate40%9.3%
Cart Rate20%29.7%
Search Rate10%78%
Turn Efficiency15%6%
Token Efficiency15%17%
22% of sessions failed with errors.

Top Stores

StoreCheckout %Cart %
••••••••••••••••••••••••••••66.7%66.7%
everlane.com60%60%
••••••••••••••••••••••••••••••••33.3%33.3%
•••••••••••••••••••••••••••••0%0%
craze.wtf0%0%
goswingofficial.com0%50%
•••••••••••••••••0%0%
teveo.com0%50%
westwing.de0%33.3%
••••••••••••••••••••••••••••••••0%100%

Known Issues

Non-Functionalhigh

92.9% tool error rate — almost every tool call fails.

Schema Incompatibilityhigh

Cannot correctly format MCP tool arguments. Not usable for shopping agent tasks.

Token Usage

Avg per Session70,680
Fleet Average67,228
vs Fleet+5%
Median (p50)29,882
p90152,402
Prompt / Completion split
Prompt 67,948 (96%) Completion 2,732 (4%)
Daily Average Tokens (last 30 days)
Distribution range
Min0p2515,531p5029,882p7577,176p90152,402Max721,744

Test o4-mini (Reasoning) on Your Store

Run a live agent session to see how o4-mini (Reasoning) handles your store's checkout flow end-to-end.

Run Agent Session

Compare With Other Models

MetaMetaLlama 3.3 70B39.3%GoogleGoogleGemini 3.5 Flash38.7%AnthropicAnthropicClaude Sonnet 4.636.5%xAIxAIGrok 4.326.3%GoogleGoogleGemini 2.5 Flash25.7%AnthropicAnthropicClaude Opus 4.823.3%DeepSeekDeepSeekDeepSeek V3.222.1%GoogleGoogleGemini 3.1 Pro20.9%GoogleGoogleGemini 2.5 Pro20%OpenAIOpenAIGPT-4o14.5%xAIxAIGrok 3 Mini (Reasoning)11.1%OpenAIOpenAIGPT-5.511%OpenAIOpenAIo4-mini (Reasoning)9.3%AnthropicAnthropicClaude Sonnet 58.6%DeepSeekDeepSeekDeepSeek V4 Pro8.6%xAIxAIGrok 4.57.7%AnthropicAnthropicClaude Opus 56.8%DeepSeekDeepSeekDeepSeek V4 Flash6.5%Moonshot AIMoonshot AIKimi K34.3%DeepSeekDeepSeekDeepSeek R1 (Reasoning)4.1%GoogleGoogleGemini 3.6 Flash3.2%MetaMetaLlama 4 Maverick2.6%AlibabaAlibabaQwen 3.7 Max0%AlibabaAlibabaQwQ 32B (Reasoning)0%
Full comparison view →