UCP PlaygroundDeepSeek R1 (Reasoning)/Models/DeepSeek R1 (Reasoning)
InspectorAgentDocs
DeepSeekDeepSeek

DeepSeek R1 (Reasoning)

DeepSeek's reasoning model with chain-of-thought. Slow (76s avg) but deliberate — the reasoning traces show it carefully planning each tool call. Moderate error rate (12.5%) partly from JSON formatting issues. Small sample size but shows promise for complex checkout flows where thinking through validation errors matters.

Avg Tokens22,859Avg Duration63.8sTurns to Checkout6.7
CheckoutCartSearchTurnsTokens3.68.338.16642
Shopping Score23/100
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%3.6%
Cart Rate20%8.3%
Search Rate10%38.1%
Turn Efficiency15%66%
Token Efficiency15%42%
61.9% of sessions failed with errors.

Top Stores

StoreCheckout %Cart %
everlane.com33.3%33.3%
allbirds.com33.3%33.3%
houseofparfum.nl20%40%
••••••••••••••••0%0%
••••••••••••••••0%0%
••••••••••••••••0%0%
••••••••••••••••0%0%
••••••••••••••••0%0%
••••••••••••••••0%0%
••••••••••••••••0%0%

Known Issues

High Latencymedium

76 seconds average — reasoning overhead adds significant time.

JSON Formattingmedium

12.5% error rate partly from JSON formatting issues in tool arguments.

Token Usage

Avg per Session22,859
Fleet Average74,927
vs Fleet−69%
Median (p50)14,328
p9051,089
Prompt / Completion split
Prompt 21,626 (95%) Completion 1,233 (5%)
Daily Average Tokens (last 30 days)
Distribution range
Min0p257,425p5014,328p7531,881p9051,089Max116,012

Test DeepSeek R1 (Reasoning) on Your Store

Run a live agent session to see how DeepSeek R1 (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 Flash38.1%AnthropicAnthropicClaude Sonnet 4.634.7%xAIxAIGrok 4.326.3%GoogleGoogleGemini 2.5 Flash23.3%AnthropicAnthropicClaude Opus 4.822.3%AlibabaAlibabaQwen 3.8 Flash22.2%DeepSeekDeepSeekDeepSeek V3.220.7%GoogleGoogleGemini 2.5 Pro19.8%GoogleGoogleGemini 3.1 Pro19.6%OpenAIOpenAIGPT-4o15%DeepSeekDeepSeekDeepSeek V4.1 Flash14.3%xAIxAIGrok 3 Mini (Reasoning)11.1%OpenAIOpenAIGPT-5.510.2%OpenAIOpenAIo4-mini (Reasoning)9.1%AnthropicAnthropicClaude Opus 59%xAIxAIGrok 4.58.1%AnthropicAnthropicClaude Sonnet 57.7%DeepSeekDeepSeekDeepSeek V4 Pro7.7%DeepSeekDeepSeekDeepSeek V4 Flash7.5%GoogleGoogleGemini 3.8 Flash6.7%DeepSeekDeepSeekDeepSeek R1 (Reasoning)3.6%Moonshot AIMoonshot AIKimi K33.1%AlibabaAlibabaQwen 3.7 Max2.6%MetaMetaLlama 4 Maverick2.3%GoogleGoogleGemini 3.6 Flash1.1%GoogleGoogleGemini 3.7 Flash0%OpenAIOpenAIGPT-6 Luna0%MetaMetaMuse Spark 1.30%AlibabaAlibabaQwQ 32B (Reasoning)0%
Full comparison view →