Chat / LLM
Generate text responses. OpenAI-compatible format. Supports streaming.
48 models available
https://api.nextgenrelay.com/api/v1/chat/completions| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| model | string | required | — | Model ID (see list below) |
| messages | array | required | — | [{"role":"user","content":"Hello"}]. Roles: system, user, assistant |
| stream | boolean | optional | false | Enable SSE streaming |
| temperature | number | optional | model default | Sampling temperature (0–2) |
| max_tokens | integer | optional | 8192 (Anthropic) | Max output tokens |
cURL Example
curl -X POST https://api.nextgenrelay.com/api/v1/chat/completions \
-H "Authorization: Bearer sk_live_xxx" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4o",
"messages": [
{"role": "system", "content": "You are helpful."},
{"role": "user", "content": "Hello!"}
],
"stream": false,
"temperature": 0.7
}'Response
{
"id": "chatcmpl-xxx",
"model": "gpt-4o",
"choices": [{
"message": {"role": "assistant", "content": "Hello! How can I help?"},
"finish_reason": "stop"
}],
"usage": {"prompt_tokens": 10, "completion_tokens": 15, "total_tokens": 25}
}Available Models
Alibaba
Qwen 3.7 Max
6crAlibaba's flagship model
Qwen 3.7 Plus
4crBalanced performance and speed
Qwen 3.6 Plus
4crHigh-quality Qwen model
Qwen 3.6 Flash
2crFast and efficient
Qwen 3.5 Flash
2crUltra-fast Chinese-English model
Anthropic
Claude Opus 4.8
18crMost powerful Claude for complex tasks
Claude Sonnet 4.6
8crBest balance of intelligence and speed
Claude Haiku 4.5
3crLightning-fast responses
Claude Fable 5
12crAnthropic's creative storytelling model
ByteDance
Doubao Seed 2.0 Pro
6crByteDance's most capable model
Doubao Seed 2.0 Mini
2crLightweight and fast
DeepSeek
DeepSeek V4 Pro
6crTop-tier Chinese AI with strong reasoning
DeepSeek V4 Flash
2crFast and affordable
DeepSeek R1
8crOpen-source reasoning champion
DeepSeek V3.2
4crLatest DeepSeek general model
Gemini 3.5 Flash
5crLatest Gemini with huge context
Gemini 2.5 Pro
12crGoogle's most capable reasoning model
Gemini 2.5 Flash
3crFast and efficient multimodal model
Microsoft
Phi-4 Reasoning
3crSmall but powerful reasoning model
MiniMax
MiniMax M3
5crLatest MiniMax flagship
MiniMax M2.7 Fast
3crHigh-speed generation
Mistral
Mistral Large
6crEuropean AI flagship
Devstral Medium
4crOptimized for developers
Moonshot
Kimi K2.7 Code
8crElite coding model
OpenAI
GPT-5.4
15crMost advanced OpenAI model with deep reasoning
GPT-5.4 Mini
8crBalanced performance and speed
GPT-5.4 Nano
3crUltra-fast responses for simple tasks
GPT-5.2
12crOpenAI's latest high-performance model
GPT-5.1
10crBalanced performance and efficiency
GPT-5
10crGPT-5 base model
GPT-5 Mini
3crFast and efficient GPT-5 variant
GPT-5 Nano
1crUltra-low-cost GPT-5 variant
GPT-4.1
8crLatest GPT-4 series model
GPT-4.1 Mini
3crEfficient GPT-4.1 variant
GPT-4.1 Nano
1crBudget GPT-4.1 variant
GPT-4o
5crVersatile multimodal workhorse
GPT-4o Mini
2crCost-efficient for everyday tasks
o3
20crAdvanced reasoning and problem solving
o4 Mini
8crEfficient reasoning model
Perplexity
Sonar Pro
6crAI-powered search with citations
Sonar Reasoning Pro
10crSearch + deep reasoning
StepFun
Step 3
4crStepFun's general-purpose model
Step 3.7 Flash
2crFast StepFun model
Tencent
Hunyuan T1
4crTencent's reasoning model
xAI
Grok 4.3
8crWitty, real-time, uncensored
Grok 3
5crFast and capable
Zhipu
GLM-5.2
8crZhipu's latest multimodal flagship
GLM-4.7
3crReliable and affordable