mirror of
https://git.openapi.site/https://github.com/desirecore/config-center.git
synced 2026-10-21 01:24:43 +08:00
* docs: split model sources and consolidate project instructions * docs: normalize empty model navigation footer
5.7 KiB
5.7 KiB
qwen3.8-flash:模型来源与接入记录
- 精确模型 ID:
qwen3.8-flash。 - 核验日期:2026-10-03;本次只迁移记录,没有重新核验或提高状态。
适用接入面
- compute/providers/dashscope.json
- compute/coding-plans/dashscope-token-plan.json
- compute/model-specs/qwen.json
字段级来源记录
各接入面分开记录;价格为 inputPrice/outputPrice 快照,具体单位和限制看配置及引用官网。partial 只确认已列字段,不代表全部参数、平台可用性或账号调用已验收。
compute/providers/dashscope.json
当前配置详情(不是全部官网确认值;routing 是本仓策略):
{
"contextWindow": 1000000,
"maxOutputTokens": 131072,
"serviceType": [
"chat",
"vision"
],
"capabilities": [
"chat",
"reasoning",
"code",
"multilingual",
"long_context",
"tool_use",
"agent",
"vision",
"video_understanding"
],
"defaultTemperature": 0.6,
"defaultTopP": 0.95,
"inputPrice": 0.8,
"outputPrice": 2.7,
"extra.reasoning": {
"supportedEfforts": [
"low",
"medium",
"xhigh"
],
"defaultEffort": "xhigh"
},
"extra.cachedInputPrice": 0.1,
"extra.pricingNotes": "CNY per 1M tokens, Beijing region standard price."
}
| 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
|---|---|---|---|---|---|---|
qwen3.8-flash |
1000000 / 131072 | CNY:0.8 / 2.7 | modelName→qwen-models |
qwen-models | partial | 0cd69defad23918637fcf9c7f319a5b6a6a32d3fd224a6fe120b718044e296a3 |
compute/coding-plans/dashscope-token-plan.json
当前配置详情(不是全部官网确认值;routing 是本仓策略):
{
"contextWindow": 1000000,
"maxOutputTokens": 131072,
"serviceType": [
"chat",
"reasoning",
"vision"
],
"capabilities": [
"chat",
"reasoning",
"deep_thinking",
"code",
"vision",
"tool_use",
"agent",
"long_context",
"fast"
],
"defaultTemperature": 0.6,
"extra.reasoning": {
"supportedEfforts": [
"low",
"medium",
"xhigh"
],
"defaultEffort": "xhigh"
}
}
| 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
|---|---|---|---|---|---|---|
qwen3.8-flash |
1000000 / 131072 | 套餐:未声明 / 未声明 | modelName→qwen-models |
qwen-models | partial | c970d11813e282b030abd43126cccc17a142c920d66eac7e5d94da9936b334d3 |
compute/model-specs/qwen.json
当前配置详情(不是全部官网确认值;routing 是本仓策略):
{
"spec.contextWindow": 1000000,
"spec.maxOutputTokens": 131072,
"spec.serviceType": [
"chat",
"reasoning",
"vision"
],
"spec.capabilities": [
"chat",
"reasoning",
"deep_thinking",
"code",
"multilingual",
"tool_use",
"long_context",
"agent",
"vision",
"video_understanding",
"fast"
],
"spec.supportsReasoning": true,
"spec.defaultTemperature": 0.6,
"spec.defaultTopP": 0.95,
"spec.releasedAt": "2026-08-26",
"routing.tier": "lightweight",
"routing.routingPriority": 40,
"routing.eligibleForAgent": true,
"routing.reasoning": {
"supportedModes": [
"auto",
"off",
"minimal",
"low",
"medium",
"high",
"xhigh",
"max"
],
"defaultMode": "xhigh"
}
}
| 模型 ID | 上下文 / 输出 | 计价 | 已核字段→来源 | 来源入口 | 状态 | 数据指纹 |
|---|---|---|---|---|---|---|
qwen3.8-flash |
1000000 / 131072 | 非计价主数据 | id→qwen-models |
qwen-models | partial | d5845de5effb40061fbf50bf3aac30e560dd972d0ec2ad3e2ed58b60fb825026 |
下次更新核查
- 对照官网核查精确 ID/别名、是否仍列出、上下文/输入/输出限制及单位。
- 分别核查推理档位、默认值、是否可关闭思考、采样和强制工具选择限制、多模态输入/输出。
- 按本接入面核对价格、缓存、阶梯/峰谷、套餐支持;不能把其他平台参数直接复制。
- 只更新真正复核的字段与引用来源;保留其余待核实项及历史记录。
- 同步对应 canonical JSON、回归和模型指纹,再运行来源校验。
{
"formatVersion": 1,
"supplier": "alibaba",
"checkedAt": "2026-10-03",
"sourceCatalog": "../SOURCES.md"
}