Omni-modal reasoner that reads text, images, audio, and video and answers in text, with a 1M token context, thinking, and native web search.
Alibaba CloudText Generation1.01M contextReleased Sep 17, 2026SingaporeProprietary EndpointNewOmni-modal reasoner that reads text, images, audio, and video and answers in text, with a 1M token context, thinking, and native web search.
Supports text, image, audio, and video input, thinking mode with enable_thinking and thinking_budget up to 262144 tokens, function calling, structured output including strict JSON Schema, and the tool_web_search built-in tool. Thinking tokens are billed as output tokens.
Also known as Alibaba Cloud Qwen3.8 Omni Flash
qwen3-8-omni-flash/v1/chat/completionsPOST/v1/responsesPOST/v1/messagesPOST/v1beta/models/qwen3-8-omni-flash:generateContentqwen3.8-omni-flashLive pay-as-you-go rates from the EmpirioLabs catalog. You are billed only for what you use, with no monthly minimum.
Qwen3.8 Omni Flash serves the OpenAI-compatible Chat Completions API. Point any OpenAI SDK at https://api.empiriolabs.ai/v1 with your EmpirioLabs API key and use the model id qwen3-8-omni-flash. Get an API key from the EmpirioLabs dashboard.
curl https://api.empiriolabs.ai/v1/chat/completions \
-H "Authorization: Bearer $EMPIRIOLABS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen3-8-omni-flash",
"messages": [
{"role": "user", "content": "Write a haiku about the ocean."}
]
}'from openai import OpenAI
client = OpenAI(
base_url="https://api.empiriolabs.ai/v1",
api_key="YOUR_EMPIRIOLABS_API_KEY",
)
response = client.chat.completions.create(
model="qwen3-8-omni-flash",
messages=[{"role": "user", "content": "Write a haiku about the ocean."}],
)
print(response.choices[0].message.content)Request parameters supported by the Qwen3.8 Omni Flash API on EmpirioLabs. Defaults apply when a field is omitted.
| Parameter | Type | Default | Range / values | Description |
|---|---|---|---|---|
| temperature | number | 0.7 | 0 to 1.99 | Sampling temperature. 0 is deterministic and 1.99 is maximum randomness. |
| top_p | number | 0.9 | 0 to 1 | Nucleus sampling probability mass. Lower values make outputs more focused. |
| max_tokens | number | - | 1 to 131072 | Maximum output tokens. When omitted, the model's own default applies. |
| stop | string | - | - | Up to 4 strings where the model will stop generating further tokens. |
| enable_thinking | boolean | true | - | Enable reasoning before answering. A request that forces a tool call runs without reasoning unless it sets enable_thinking. |
| reasoning_effort | enum | - | none, low, medium, high, max | Reasoning effort level. none disables thinking. low, medium, high, and max set bounded thinking budgets sized to the selected model. |
| thinking_budget | number | - | 1 to 262144 | Maximum tokens reserved for reasoning when thinking is enabled. When omitted, the model decides how long to reason. |
| preserve_thinking | boolean | false | - | Include the reasoning of earlier assistant turns in the model's input on /v1/chat/completions. Return each turn's reasoning unchanged in that assistant message's... |
| seed | integer | - | 0 to 2147483647 | Random seed for more reproducible sampling. Reusing a seed with the same request returns similar, not always identical, output. |
| top_k | integer | - | 1 to 100 | Sample only from the K most likely tokens. When omitted, the model's own default applies. |
| presence_penalty | number | - | -2 to 2 | Penalty for tokens that already appeared in the generated text. Positive values encourage new topics. When omitted, the model's own default applies. |
| repetition_penalty | number | - | 0.1 to 2 | Higher values discourage repeated phrasing, and 1 applies no penalty. When omitted, the model's own default applies. Not available when a built-in tool is on, or on... |
| n | integer | 1 | 1 to 4 | Number of answers to return. Above 1 runs only on /v1/chat/completions, with enable_thinking false and without tools, web search or logprobs. The output tokens of... |
| logprobs | boolean | false | - | Return the log probabilities of the answer tokens, not the reasoning, on /v1/chat/completions. Not available when a built-in tool is on. |
Text, image, audio, and video input are supported, and the model replies in text. Thinking is on by default; use enable_thinking to turn it off and thinking_budget to cap it at up to 262144 tokens. Thinking tokens are billed as output tokens. Web search is an optional built-in tool exposed through tool_web_search and adds $0.02 for each invoked call; a single request can invoke it more than once, and each invoked call is billed. For audio input, pass the file as a URL or a base64 data URI. A video's soundtrack is billed as audio tokens alongside its video tokens.
When this model invokes built-in tools inside a single request, the response carries a normalized usage.tool_usage map alongside the token counts:
"usage": {
"prompt_tokens": 123,
"completion_tokens": 456,
"cost_usd": 0.0042,
"tool_usage": {"web_search": 3}
}Tool counts are already factored into cost_usd and are surfaced for transparency so you can audit per-tool billing. The field is omitted when no tools were invoked.
On EmpirioLabs, Qwen3.8 Omni Flash is billed pay as you go: Input $0.30 per 1M prompt tokens; Output $0.94 per 1M generated tokens; Web search $0.02 per call when invoked. The live rate card on this page always matches what the API charges.
Qwen3.8 Omni Flash supports a 1.01M-token context window with up to 131,072 output tokens per response.
Yes. Qwen3.8 Omni Flash serves the OpenAI-compatible Chat Completions API, so existing OpenAI SDKs work by pointing base_url at https://api.empiriolabs.ai/v1 and setting the model id to qwen3-8-omni-flash.
Yes. The EmpirioLabs playground runs Qwen3.8 Omni Flash in the browser with the same parameters the API exposes, so you can test prompts before writing code.
Create an EmpirioLabs account, then generate a key under API Keys in the dashboard. Billing is pay-as-you-go credits, so you only pay for the requests you make.
Check out our pricing or reach out if you want your own model deployed on our stack.
