Skip to content
Abhigya AI

One endpoint.Every model we host.

Autoscaling, scale to zero, and per-token pricing - no GPUs to reserve and no capacity to plan.

OpenAI-compatible · scale to zero · per-token billing
$ pip install openai
quickstart.py
from openai import OpenAI

client = OpenAI(
    base_url = "https://api.abhigya.ai/v1",
    api_key = "<ABHIGYA_API_KEY>",
)

stream = client.chat.completions.create(
    model    = "qwen/qwen2.5-7b-instruct",
    messages = [{"role": "user",
                 "content": "Explain serverless inference."}],
    stream   = True,
)
What you get

Production behaviour, by default.

Autoscaling

Capacity follows the queue, not a forecast. Spikes absorb without a ticket.

Scale to zero

An idle endpoint costs nothing.

Streaming and tools

Token streaming, tool calling, and structured output on one endpoint.

Model catalog

  • DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash · 1M contextIn ₹20.00 · Out ₹60.00 · Cached ₹0.60 /M
  • Kimi K3moonshot/kimi-k3 · 1M contextIn ₹285.00 · Out ₹1,425.00 · Cached ₹28.50 /M
  • GLM 5.3 Flashzai/glm-5.3-flash · 1M contextIn ₹15.00 · Out ₹50.00 · Cached ₹3.00 /M
  • GLM 5.3zai/glm-5.3 · 1M contextIn ₹90.00 · Out ₹400.00 · Cached ₹20.00 /M
  • Qwen 3.8 Flash-Nextqwen/qwen3.8-flash-next · 256K contextIn ₹15.00 · Out ₹47.00 · Cached ₹1.60 /M
  • Qwen 3.8 27Bqwen/qwen3.8-27b · 256K contextIn ₹20.00 · Out ₹250.00 · Cached ₹5.00 /M
  • Qwen2.5 7B Instructqwen/qwen2.5-7b-instruct · 128K contextIn ₹10.00 · Out ₹20.00 /M
  • Gemma 4 31Bgoogle/gemma-4-31b-it · 256K contextIn ₹20.00 · Out ₹40.00 /M
  • Gemma 4 26B A4Bgoogle/gemma-4-26b-a4b-it · 256K contextIn ₹7.00 · Out ₹34.00 /M