Model Configuration
This article covers the models, vendor configurations, data center availability, and pricing for Zia Agents. If you're setting up your first agent and wondering which model to pick or what it's going to cost, this is where you'll find those answers.
Supported vendors and models
When you create an agent, one of the first things you configure is the vendor and model. Zia Agents supports two vendor configuration types.
ZKS (Zoho Key System)
Zoho manages these models for you, so there's no API key to set up. They're available out of the box.
| Vendor | Models |
| OpenAI (through Zoho's keys) | GPT-4o, GPT-4o Mini, GPT-4.1, GPT-4.1 Mini, GPT-5.1, GPT-5.4, GPT-5.4 Mini, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna |
| Zoho-hosted (on Zoho's infrastructure) | Qwen 14B, Qwen 3.5 35B MoE, Qwen 3.5 122B MoE, GLM 4.7 Flash, GLM 5 |
BYOK (Bring Your Own Key)
Connect your own API key from a supported vendor and use their models directly. Zia Agents doesn't charge anything for this. You pay the vendor on their billing cycle, and the platform routes the requests. BYOK supports six vendors:
| Vendor | Models |
| OpenAI | All OpenAI models available through ZKS, plus GPT-6 Astra, GPT-6 Sol, GPT-6 Luna, GPT-4.1 (2025-04-14), GPT-4.1 Mini (2025-04-14), GPT-4o (2024-08-06), GPT-4o Mini (2024-07-18), GPT-4 Turbo, GPT-3.5 Turbo |
| Anthropic | Claude Haiku 4.5, Claude Sonnet 4.5, Claude Sonnet 4.6, Claude Sonnet 5, Claude Opus 4.5, Claude Opus 4.6, Claude Opus 4.7, Claude Opus 4.8, Claude Opus 5 |
| Google Gemini | Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash Lite, Gemini 3.1 Pro Preview, Gemini 3.1 Flash Lite Preview, Gemini 3.5 Flash |
| DeepSeek | DeepSeek Chat, DeepSeek V4 Flash, DeepSeek V4 Pro |
| Cohere | Command A, Command R, Command R+ |
| Siliconflow | Qwen 3 Coder 30B A3B Instruct, Qwen 3.6 35B A3B, Qwen 3.8 2.4T A95B, GLM 5.2, GLM 5.3, GLM 5.3 Flash, GLM 5V Turbo, DeepSeek V3, DeepSeek V4 Flash, Kimi K3, MiniMax M3 |
The choice between ZKS and BYOK usually comes down to two things: whether your organization has an existing API arrangement with a vendor, and whether data residency matters to you. With Zoho-hosted models under ZKS, your data stays within Zoho's infrastructure. With BYOK, data is processed by whichever vendor you connect to.
Retired models: These models have been retired and are no longer available in Zia Agents. If an agent still uses one of them, switch it to a current model.
- OpenAI: GPT-4 1106 Preview
- Anthropic: Claude 3 Opus, Claude 3 Haiku, Claude 3.5 Sonnet, Claude 3.7 Sonnet, Claude Sonnet 4, Claude Opus 4
- Google Gemini: Gemini 2.0 Flash, Gemini 2.0 Flash Lite
- Zoho (ZKS): Qwen 30B MoE (replaced by Qwen 3.5 35B MoE)
- Cohere: Command A Translate
Model availability by data center
| US | EU | CA | IN | JP | CN | AU | |
| OpenAI (BYOK) | ✓ | ✓ | ✓ | ✓ | ✓ | - | ✓ |
| Anthropic (BYOK) | ✓ | ✓ | ✓ | ✓ | ✓ | - | ✓ |
| Gemini (BYOK) | ✓ | ✓ | ✓ | ✓ | ✓ | - | ✓ |
| DeepSeek (BYOK) | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Cohere (BYOK) | ✓ | ✓ | ✓ | ✓ | ✓ | - | ✓ |
| Siliconflow (BYOK) | - | - | - | - | - | ✓ | - |
| OpenAI (ZKS) | ✓ | - | ✓ | ✓ | ✓ | - | ✓ |
| Qwen 14B (ZKS) | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Qwen 3.5 35B MoE (ZKS) | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Qwen 3.5 122B MoE (ZKS) | ✓ | ✓ | ✓ | ✓ | - | ✓ | ✓ |
| GLM 4.7 Flash (ZKS) | ✓ | ✓ | - | ✓ | ✓ | - | ✓ |
| GLM 5 (ZKS) | ✓ | - | ✓ | ✓ | ✓ | - | ✓ |
Notes:
- Availability depends on the vendor, the configuration type (ZKS or BYOK), and the data center of your organization.
- If a model doesn't show up in your agent creation screen, it's likely not available for your data center. Check this table or contact support for the latest availability.
- A dash (-) means the vendor or model isn't offered in that data center yet. Support for more data centers, including SG, UK, and SA, is on the way.
Pricing and credits
Zia Agents is a free-to-use platform. The agent builder, the deployment infrastructure, and BYOK integration all come at no cost. What you pay for, if anything, depends entirely on which models you use and how much.
There are three billing scenarios to understand:
- BYOK (Bring Your Own Key) is entirely free on the Zia Agents side. You connect your own API key, the vendor bills you directly based on their rates, and Zia Agents doesn't add a markup. Nothing to track here.
- Zoho-hosted models through ZKS (the Qwen and GLM families) come with a monthly free token allowance. For most teams, this covers typical usage entirely. You only start paying if you exceed the allowance. These models are organized into two tiers:
| Standard Tier | Pro Tier | |
| Models | Qwen 14B, Qwen 3.5 35B MoE, GLM 4.7 Flash | GLM 5 |
| Monthly free allowance | 30 million tokens | 20 million tokens |
| Rate after free allowance | $1.00 per 1M tokens | $3.00 per 1M tokens |
| Credit consumption | 1 credit per 1,000 tokens | 3 credits per 1,000 tokens |
Local taxes (VAT, GST, etc.) will be charged in addition to the prices mentioned.
Note: GLM 5, the Pro tier model, is available on a request basis. The data center table shows where it's offered.
External models through ZKS (OpenAI models accessed using Zoho's keys, not your own) do not have a free allowance. Usage is billed at exact vendor rates with no markup, and credits are deducted from your wallet from the first token.
The difference between this and BYOK is who manages the API key. With BYOK, you handle the key and the vendor bills you. With external models through ZKS, Zoho manages the key and deducts credits from your Zia Agents wallet instead.
How credits work
Zia Agents uses a prepaid credit system to keep costs predictable.
The exchange rate is 1 USD = 1,000 AI Credits. You maintain a prepaid wallet, and credits are only consumed once you've exhausted your monthly free allowance on Zoho-hosted models or when using external models through ZKS. Purchased credits roll over month to month. The free monthly token allowance does not roll over and resets on the 1st of each month.
To view your current credit balance, purchase credits, or see usage breakdowns by model and agent, go to Settings > Plan & Billing.
What this looks like in practice
- A team using only Zoho-hosted Standard models. They run Qwen 3.5 35B MoE for internal support bots and consume 25 million tokens in a month. Cost: zero. They're within the 30 million free Standard allowance.
- A high-volume production team. They run GLM 5 (Pro Tier) for a complex autonomous agent and consume 25 million tokens. The first 20 million are covered by the free Pro allowance. The remaining 5 million cost 15,000 AI Credits, which is $15.00.
- A team using OpenAI models through ZKS. They use GPT-4.1 through Zoho's keys. Usage is deducted from their wallet at exact vendor rates from the first token. No free allowance applies.
- A team using BYOK. They connect their own Anthropic API key and use Claude Sonnet 4.6. Cost on Zia Agents: zero. Anthropic bills them directly.