Documentation Index Fetch the complete documentation index at: https://docs.ninjachat.ai/llms.txt
Use this file to discover all available pages before exploring further.
Request curl -X POST https://ninjachat.ai/api/v1/chat \ -H "Authorization: Bearer nj_sk_YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "auto", "messages": [{"role": "user", "content": "Solve: x² + 5x + 6 = 0"}], "include_routing": true }'
Response { "model" : "o3-mini" , "choices" : [{ "message" : { "role" : "assistant" , "content" : "Factoring: x² + 5x + 6 = (x + 2)(x + 3) = 0 \n Solutions: x = -2 and x = -3" }, "finish_reason" : "stop" }], "routing" : { "requested" : "auto" , "resolved" : "o3-mini" , "task_type" : "math" , "reasoning" : "Detected task type: math" }, "cost" : { "this_request" : "$0.006" } }
Add "include_routing": true to see which model was chosen and why.
The four variants Model ID Optimizes for Best when… autoQuality + speed balance You want the best model without thinking about it auto-fastLowest latency Real-time apps, chatbots, low-latency pipelines auto-cheapLowest cost High-volume jobs, batch processing, cost-sensitive apps auto-qualityHighest quality Critical decisions, best possible output
How task detection works NinjaChat analyzes your last 3 user messages to detect the task type:
Task type Detected keywords auto routes tocodefunction, debug, implement, algorithm, TypeScript, SQL… claude-sonnet-4.6mathequation, solve, calculate, integral, probability… o3-minicreativewrite, story, poem, imagine, fiction, lyrics… gemini-3.1-proanalysisanalyze, compare, evaluate, research, summarize… gpt-5quickShort prompts under 80 chars, “what is”, “define”… gemini-3-flashgeneralEverything else gpt-5
Full routing table auto (balanced)
auto-fast
auto-cheap
auto-quality
Task Model code claude-sonnet-4.6math o3-minicreative gemini-3.1-proanalysis gpt-5quick gemini-3-flashgeneral gpt-5
Task Model code claude-haiku-4.5math gpt-5-minicreative gemini-3-flashanalysis gpt-5-miniquick gpt-5-minigeneral gpt-5-mini
Task Model code deepseek-v3math qwq-32bcreative gemini-3-flashanalysis deepseek-v3quick gemini-2.5-flashgeneral llama-4-maverick
Task Model code claude-opus-4.6math o3-minicreative gemini-3.1-proanalysis claude-opus-4.6quick claude-opus-4.6general claude-opus-4.6
Billing Auto variants are billed at the resolved model’s rate . If auto routes to o3-mini, you pay 0.006. I f i t r o u t e s t o ‘ c l a u d e − s o n n e t − 4.6 ‘ , y o u p a y 0.006. If it routes to `claude-sonnet-4.6`, you pay 0.006. I f i t ro u t es t o ‘ c l a u d e − so nn e t − 4.6‘ , yo u p a y routing field always shows the cost-incurring model.
Parameters Parameter Type Default Description include_routingboolean falseInclude routing object in response. budget_centsnumber — Override with a cost ceiling. See Budget Routing .
Code examples Python
Node.js
cURL — auto-cheap
import requests, os r = requests.post( "https://ninjachat.ai/api/v1/chat" , headers = { "Authorization" : f "Bearer { os.environ[ 'NINJACHAT_API_KEY' ] } " }, json = { "model" : "auto" , "messages" : [{ "role" : "user" , "content" : "Write a merge sort in Python" }], "include_routing" : True , } ) data = r.json() print (data[ "choices" ][ 0 ][ "message" ][ "content" ]) print ( "Routed to:" , data[ "routing" ][ "resolved" ]) # claude-sonnet-4.6 print ( "Task type:" , data[ "routing" ][ "task_type" ]) # code
Manual model selection If you prefer explicit control over which model runs, here’s a quick reference:
models = { "classify" : "deepseek-v3" , # $0.003 — simple tasks "chat" : "gpt-5" , # $0.006 — general use "code" : "claude-sonnet-4.6" , # $0.015 — when quality matters "fast" : "gemini-2.5-flash" , # $0.003 — lowest latency }
See Models for the full list with pricing and recommendations. 
