Open for developersOpenAI-compatible

One API. Intelligent model routing.

Access AtlasFlux Nenas Flash through a single OpenAI-compatible API with reasoning, multimodal understanding, web search and automatic model routing.

No monthly commitmentPrepaid balance in RMUsage-based billing
bash
curl https://api.atlasflux.my/v1/responses \
  -H "Authorization: Bearer $ATLASFLUX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "atlasflux/nenas-flash",
    "input": "Explain this architecture",
    "reasoning": {
      "effort": "medium"
    }
  }'
json
{
  "id": "resp_8f2k3m9x",
  "model": "atlasflux/nenas-flash",
  "output": "This architecture uses a single API gateway...",
  "usage": {
    "input_tokens": 1842,
    "output_tokens": 512,
    "reasoning_tokens": 640
  },
  "cost_myr": 0.0192
}

Everything you need to ship with AI

A developer platform built around one stable model ID, transparent pricing and the control surfaces you expect.

Intelligent model routing

Each request is routed to the best upstream model for the task. One model ID stays stable while routing adapts underneath.

Adjustable reasoning

Control reasoning effort per request with low, medium and high levels to balance quality and cost.

Built-in web search

Ground answers in current sources with configurable search depth, result limits and content extraction.

Multimodal input

Send text, images, audio and documents through the same endpoint for understanding beyond plain text.

OpenAI-compatible API

Drop-in compatible with existing OpenAI SDKs. No new client libraries required to get started.

Usage-based billing in MYR

One prepaid API balance in Malaysian Ringgit. No invoices, no monthly plans, no minimums.

Streaming responses

Token-by-token streaming for responsive chat experiences, with the same cost accounting as batch calls.

Detailed usage analytics

Understand spend, tokens, searches, latency and errors down to the request level in the dashboard.

The AtlasFlux model

AtlasFlux exposes one primary model ID. Behind it, requests are optimised for the task at hand so you get good results without managing providers.

atlasflux/nenas-flash

Nenas Flash

The AtlasFlux model. One ID for reasoning, multimodal input, web search and structured output.

  • Adjustable reasoning effort
  • Built-in web search
  • Multimodal input
  • Streaming responses
  • Tool calling
  • Structured output

Transparent routing

AtlasFlux routes each request to the most suitable upstream model. The public model ID never changes, even if the underlying routing or upstream availability is adjusted to improve quality and reliability.

InputRM5 per 1M tokens
OutputRM25 per 1M tokens
Web searchfrom RM0.05 per search

Simple, prepaid pricing in MYR

All usage is deducted from a single prepaid API balance. Top up when you need to, with no monthly commitment and no minimum spend.

Rate card

  • Input

    Prompt and system tokens

    RM5

    per 1M tokens

  • Output

    Generated tokens

    RM25

    per 1M tokens

  • Web search

    From, depending on depth

    RM0.05

    per search

  • Reasoning

    Reasoning tokens billed at the output rate

    Output rate

    per 1M tokens

Reasoning tokens are billed at the output rate. Search results above 10 and extracted content pages are billed as additional usage.

Top-up options

Popular: RM100
RM10Credit
RM25Credit
RM50Credit
RM100Most popular
RM250Credit
RM500Credit

Topping up adds prepaid credit to your API balance. There are no plans to buy and no expiring credits.

Start building with AtlasFlux today

Create an account, generate your first API key and make a request in under five minutes. Your balance is prepaid and always visible.