Grok

AI

xAI models — chat, web/X search, code execution, image generation/editing, and text-to-speech.

83.3
out of 100

Moderate accuracy (72%) — some responses were incomplete or partially correct. Very reliable — 9/10 requests got a response. Median response time: 4623ms (p95: 30003ms).

Last tested: 3/26/2026, 3:01:45 PM

What AI models say

GPT-4o-mini

The API offers a diverse range of functionalities, from chat and web search to TTS and image generation, but the inconsistent performance and high latency detract significantly from the overall experience. While interactive tasks like chatting with empty content performed surprisingly well, the web search feature was notably unreliable and painfully slow, leaving much to be desired in usability and efficiency.

Haiku

Grok's core chat and code execution work reliably, but web search is broken—it times out consistently—and the whole platform feels half-baked for a paid service claiming multimodal capabilities. Skip it if you need search; use it if you just need another API-accessible LLM that's slightly cheaper than competitors.

Accuracy (40%)
72.0
Reliability (30%)
90.0
Latency (20%)
87.5
Cost Efficiency (10%)
100.0

Test Examples

Real requests we sent and the responses we received.

Image generation with default grok-imagine-image model

POST /grok/image-generatetypical4724ms
100%

{"data":{"data":[{"url":"https://imgen.x.ai/xai-imgen/xai-tmp-imgen-5f206c3b-7d3e-4ba7-aedd-b0d0b3aa5ea4.jpeg","mime_type":"image/jpeg","revised_prompt":""}],"usage":{"cost_in_usd_ticks":200000000}},"

Code execution with grok-4-0709 model

POST /grok/code-executiontypical19630ms
100%

{"data":{"id":"31685422-cc9e-91af-bfac-2c4b80ef2427","text":{"format":{"type":"text"}},"user":null,"error":null,"model":"grok-4-0709","store":true,"tools":[{"type":"code_interpreter"}],"top_p":0.95,"u

Chat with temperature and top_p parameters

POST /grok/chattypical4623ms
100%

{"data":{"id":"7d9e2985-cdc2-963e-87cd-cb6af4574700","model":"grok-3-mini","usage":{"total_tokens":268,"prompt_tokens":20,"num_sources_used":0,"completion_tokens":15,"cost_in_usd_ticks":1291000,"promp

Web search with grok-4-0709 model

POST /grok/web-searchtypical30003ms
0%

This operation was aborted

Service Details

API AccessOpenAPI3rd Party
by Grok
POST /grok/chat
POST /grok/image-generate
POST /grok/image-edit
POST /grok/web-search
POST /grok/x-search
POST /grok/code-execution
POST /grok/tts
Endpoint: https://grok.mpp.paywithlocus.com
Status: active
Discovered: 3/24/2026
Tests run: 10
Last tested: 3/26/2026

Test History

3/26/2026, 3:00:18 PM
score: 72%p50: 4623ms

Embed this badge

MPPrimo rating for Grok<a href="https://mpprimo.com/service/c9062972-c3f8-49ec-8622-d2a13aa4d205"><img src="https://mpprimo.com/api/badge/c9062972-c3f8-49ec-8622-d2a13aa4d205" alt="MPPrimo rating"></a>