Explore models. Find your fit

Compare capabilities, pricing and independent benchmarks. Build with the model that fits your workload.

Gemma 4 31B

Google DeepMind
Vision
256K context

31B-parameter instruction-tuned model for text and image understanding, with tool calling and optional reasoning.

GLM 5.3

Z.ai
Reasoning and coding
1M context

Large-scale reasoning model built for complex software engineering and long-horizon agent tasks.

GLM 5.3 Flash

Z.ai
Reasoning and coding
1M context

Faster open-weight reasoning model for agents and image understanding.

DeepSeek V4.1 Flash 1017

DeepSeek
Reasoning and coding
1M context

General-purpose open-weight model for long-document analysis and agentic workflows.

Sarvam 105B

Sarvam
Indian languages
128K context

Flagship 105B Mixture-of-Experts model trained from scratch for Indian-language understanding, complex reasoning, coding and agentic workflows.

Sarvam 105B Conversations

Sarvam
Indian languages
32K context

Post-trained variant of Sarvam 105B for real-time dialogue, voice agents and multi-turn conversations across Indian languages.

Pricing table view

Questions before you go live

Compare model capabilities, context limits, token prices and the published benchmark configurations. Then test answer quality, tool use and response times on your own workload. See published benchmarks

Open a model page and select Try in playground to test it in Indus. Each model page also includes API examples and migration steps. Open the playground

If you already call models through OpenRouter, Fireworks AI or Together AI, start with that migration guide. Use the OpenAI-compatible API with your Sarvam endpoint, API key and selected model ID. Validate supported parameters, tool calls and outputs before moving live traffic.

Inference runs on Sarvam-operated GPUs in India. Requests are not forwarded to the original model developer’s overseas API. External tools connected by your application have their own processing locations.

Zero data retention is available on request. Under an agreed arrangement, prompts and responses are not retained. Operational logs are scoped separately in the terms for your deployment.

Token Factory guarantees 99.95% uptime, backed by an SLA. Share your traffic, latency targets and support requirements with the Sarvam team. Discuss SLA

From model choice to production. Bring your workload, traffic expectations and latency targets. Build on Sarvam’s serving stack in India.

From model choice to production.

Bring your workload, traffic expectations and latency targets. Build on Sarvam’s serving stack in India.

Talk to our team Read the API guide

Explore models. Find your fit

Gemma 4 31B

GLM 5.3

GLM 5.3 Flash

DeepSeek V4.1 Flash 1017

Sarvam 105B

Sarvam 105B Conversations

Questions before you go live

How do I choose the right model?

Can I try a model before integrating it?

Can I move an existing application to Token Factory?

Where is inference processed?

Is zero data retention available?

What uptime SLA is available?