qwen3.8-flash API — per-token pricing and code examples
Price per 1M tokens
Cache rates apply when the provider reports cache tokens.
90% offAbout this model
qwen3.8-flash is served through OpenAI Chat Completions, OpenAI Responses, Anthropic Messages. You pay $0.015 per 1M input tokens and $0.047 per 1M output tokens. That is 90% below the official price of $0.15 and $0.47. Cache reads cost $0.0097 per 1M tokens when the provider reports cached tokens. It is included in budget packs from $1.99 up.
Where Qwen3.8-Flash fits
Alibaba Cloud calls Qwen3.8-Flash the latest multimodal model from the Qwen family and says it shines in coding assistance, agentic workflows and visual understanding. The Qwen Team's blog adds that it is well suited for high-volume applications, tool-driven workflows, and coding and coworking assistants. According to that blog, Qwen3.8-Flash is the production version of the Qwen3.8-Flash-Next architecture, which the team describes as an early preview of the architecture used in Qwen4, served with a 1M context by default and official built-in tools.
The numbers on the model page: a 1,000,000-token context, up to 991,808 input tokens (983,616 when thinking), up to 131,072 output tokens in either mode, and a chain-of-thought that can reach 262,144 tokens. Qwen3.8-Flash takes image, text and video input and returns text. That mix makes it a practical pick for agents that look at screenshots or short screen recordings and then write code or call tools.
Thinking defaults on the Flash model
Qwen3.8-Flash is a hybrid thinking model, and Alibaba Cloud ships it with thinking enabled by default. The model writes its reasoning into reasoning_content before the answer. On the OpenAI-compatible Chat Completions API, you can turn thinking off with enable_thinking: false, which Python OpenAI SDK users pass through extra_body because it is not a standard OpenAI field.
When thinking stays on, reasoning_effort accepts low, medium and xhigh, and xhigh is the default. For high-volume jobs that is often more depth than you need, so set the effort on purpose. Do not combine reasoning_effort with thinking_budget; Alibaba Cloud says setting both causes an error.
Two parameters behave differently from what OpenAI users expect. For Qwen3.8-Flash, preserve_thinking defaults to true, so earlier reasoning should go back in reasoning_content rather than being merged into content. And on the Chat Completions API, max_tokens limits only the answer, not the thinking; use max_completion_tokens to cap everything the model writes.
{
"model": "qwen3.8-flash",
"enable_thinking": false,
"messages": [{ "role": "user", "content": "Classify this support message by product area" }]
}
This is the /v1/chat/completions format. On Alibaba Cloud's own Anthropic-compatible API, Alibaba Cloud documents thinking with enabled or disabled, and output_config.effort for depth.
What it costs on Cent AI
Every call is billed per token from your prepaid USD balance, using the four rates in the price box on this page. Since thinking is on unless you switch it off, reasoning tokens are part of most Qwen3.8-Flash requests, and Cent AI bills them as output. In multi-turn use, the earlier reasoning you send back is counted by Alibaba Cloud as input. For bulk classification or extraction, turning thinking off is the most direct way to keep cost per request down.
Implicit context caching is automatic and cannot be turned off; Alibaba Cloud does not guarantee a hit. Explicit caching with cache_control of type ephemeral needs at least 1,024 tokens and lasts 5 minutes from the last hit. Cent AI applies the cache rates only to cached tokens Alibaba Cloud reports, such as prompt_tokens_details.cached_tokens.
Using it from your tools
Qwen3.8-Flash is available on /v1/chat/completions, /v1/responses and /v1/messages. Cent AI does not change your parameters silently. Alibaba Cloud lists tool calling, structured output and partial mode as supported, so function-calling agents and JSON extraction pipelines can use it directly.
Specification
- Context window
- —
- Input and output
- Text
- Endpoints
- /v1/chat/completions, /v1/responses, /v1/messages
- Smallest pack
- $1.99
Latency and error rate per model are on the status page. status.cent-ai.dev
What a request costs
Open the full calculatorCost per request
Estimate
qwen3.8-flashAn estimate at price version 5. Your bill comes from the usage the provider reports for each request.
Save $0.0232Code
curl https://api.cent-ai.dev/v1/chat/completions \
-H "Authorization: Bearer $CENT_AI_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "qwen3.8-flash", "messages": [{"role": "user", "content": "Hello"}]}'Set CENT_AI_KEY to a key from the dashboard. OpenAI SDKs use https://api.cent-ai.dev/v1; Anthropic SDKs use https://api.cent-ai.dev.
Price history
Version 5, in effect since Nov 12, 2026.
Effective
per 1M tokens
Inputold $0.015 new $0.015Cache readold $0.0015 new $0.0097Cache writeold $0.015 new $0.015Outputold $0.047 new $0.047Effective
per 1M tokens
Inputold $0.0075 new $0.015Cache readold $0.00075 new $0.0015Cache writeold $0.0075 new $0.015Outputold $0.024 new $0.047Effective
per 1M tokens
Inputold $0.0075 new $0.0075Cache readold $0.00075 new $0.00075Cache writeold $0.01 new $0.0075Outputold $0.024 new $0.024Effective
per 1M tokens
Inputold $0.0075 new $0.0075Cache readold $0.0008 new $0.00075Cache writeold $0.01 new $0.01Outputold $0.024 new $0.024
Compare with
- deepseek-v4.1-flashInput$0.015
- gpt-6-lunaInput$0.01978
- deepseek-v4-flashInput$0.015
- glm-5.3-flashInput$0.015
Questions
How much does qwen3.8-flash cost on Cent AI?
$0.015 per 1M input tokens and $0.047 per 1M output tokens. Cache reads cost $0.0097 and cache writes $0.015 per 1M tokens. Each request is rounded up once.
Is there a minimum charge?
A successful request costs at least $0.00002. A request that ends early, because the client left or the output was cut off, pays only for its tokens.
How do I pay?
Top up a prepaid USD balance with QRIS or USDT/USDC, or buy a budget pack once. Nothing renews automatically.
What if the provider does not report usage?
The request is billed from an estimate and marked estimated on the Requests page. An estimate never contains cache tokens.
Which endpoint do I call?
/v1/chat/completions, /v1/responses, /v1/messages. The OpenAI base URL is https://api.cent-ai.dev/v1.
Try qwen3.8-flash with one key
One prepaid balance for every model. Pay with QRIS or USDT/USDC.