Models
Dikaya models: GLM 5.3, Qwen3.8, DeepSeek V4.1, Pro 2, Pro and Flash. Compare exact names, API IDs, contexts, prices and demo conditions.
| Model | Context, tokens | Input | Cached input | Output |
|---|---|---|---|---|
Dikaya Flashseller-flash | 256,000 | $2 | $0.2 | $6 |
Dikaya Proseller-pro | 1,000,000 | $6 | $0.6 | $10 |
Dikaya Pro 2seller-pro-2 | 1,000,000 | $6 | $0.6 | $10 |
GLM 5.3 Abliteratedseller-glm-5-3 | 1,048,576 | $4 | $0.7 | $6 |
GLM 5.3 Flash Abliteratedseller-glm-5-3-flash | 1,048,576 | $0.5 | $0.3 | $1.6 |
Qwen3.8 Flash Next Abliteratedseller-qwen-3-8-flash-next | 262,144 | $0.36 | $0.18 | $1.2 |
DeepSeek V4.1 Flash Abliteratedseller-deepseek-v4-1-flash | 1,048,576 | $1.2 | $0.3 | $3.2 |
DeepSeek V4.1 Flash 1017 (Safeguards)seller-deepseek-v4-1-flash-1017 | 1,048,576 | $0.1 | $0.008 | $0.3 |
Dikaya Flash
Quick questions, short edits and direct image understanding.
The API identifier is seller-flash. The published context window is 256,000 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
Dikaya Pro
Large-context analysis, writing and long conversations.
The API identifier is seller-pro. The published context window is 1,000,000 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
Dikaya Pro 2
A modified version based on GLM-5.3. Code, complex tasks and creative writing. Up to 1 million tokens of context. The default demo model.
The API identifier is seller-pro-2. The published context window is 1,000,000 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
GLM 5.3 Abliterated
Modified abliterated model. Text input, low/high/max reasoning. Paid access.
The API identifier is seller-glm-5-3. The published context window is 1,048,576 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
GLM 5.3 Flash Abliterated
Modified abliterated model. Text input, low/high/max reasoning. Paid access.
The API identifier is seller-glm-5-3-flash. The published context window is 1,048,576 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
Qwen3.8 Flash Next Abliterated
Modified abliterated model. Text input, low/high/max reasoning. Paid access.
The API identifier is seller-qwen-3-8-flash-next. The published context window is 262,144 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
DeepSeek V4.1 Flash Abliterated
Modified abliterated model. Text input, low/high/max reasoning. Paid access.
The API identifier is seller-deepseek-v4-1-flash. The published context window is 1,048,576 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
DeepSeek V4.1 Flash 1017 (Safeguards)
Safeguards variant, not abliterated. Text input, low/high/max reasoning. Paid access.
The API identifier is seller-deepseek-v4-1-flash-1017. The published context window is 1,048,576 tokens. Context size is not the same as the maximum output length or the request-body byte limit.
Images: native or via text
Flash has native image input. Paid Pro and Pro 2 requests use a Flash-to-text bridge; both stages are billed. Up to eight unique images can be handled per bridged request. Embedded data/base64 image descriptions are cached per key for 24 hours; mutable HTTP URLs are not cached. Image descriptions may miss details.
Context and output limits
The paid gateway defaults to 16384 output tokens and accepts an explicit output limit up to 65536; a model may enforce a smaller native limit. JSON request bodies are limited to 16 MiB. Reasoning, when used, consumes the output budget. A large context window does not make a request free or guarantee an answer.
Practical guides
Guides on code, writing, GLM-5.3 and API integration (Russian)
Source: official Dikaya model catalogue, published rates and account/demo conditions. Pricing · Documentation
Facts reviewed: . Publisher: Dikaya.