跳至主要内容
OHLAC AI
说明中心 法律与隐私 返回首页

本页目录

  1. 积分基础
  2. 积分消耗级别
  3. 不同模型,积分消耗不同
  4. 相关链接

On this page

  1. Credit basics
  2. Credit cost levels
  3. Credits vary by model
  4. Related

このページの目次

  1. Credit basics
  2. Credit cost levels
  3. Credits vary by model
  4. Related

En esta página

  1. Credit basics
  2. Credit cost levels
  3. Credits vary by model
  4. Related

Help & Docs

积分与扣费

积分规则、预留机制与相对消耗级别说明。

版本
0.3.0
生效日期
2026 年 8 月 10 日
最后更新
2026 年 9 月 1 日

积分基础

¥1 = 500 积分(1 积分 = ¥0.002)。

积分分为购买积分与赠送积分,分别展示余额,统一进入「可用积分」。使用时先扣哪个由服务端统一管理,无需手动区分。

什么是积分预留

发送请求时,系统先预留本次可能消耗的积分。请求结束后,按服务端实际用量做最终结算,未使用的预留自动退回(预留退回)。

预留积分 ≠ 最终消费。你看到的预留金额只是上限,实际扣费通常更少。

什么是模型使用

每次成功完成的请求都会产生模型使用记录,按实际用量结算。失败、取消或无有效输出的请求释放预留,不产生扣费。

单次最低结算 1 积分。

积分消耗级别

本页不展示单个模型的精确费率明细。模型消耗以积分消耗级别作相对参考;真实计费以服务端实际用量结算为准。

低适合高频日常使用
OHLAC 1.0OHLAC 1.5GPT-5.6 Luna高性价比Gemini 3.5 Flash高性价比Gemini 3.7 Flash高性价比Grok 4.5高性价比Grok 4.6旗舰
中均衡任务,消耗适中
OHLAC ProGPT-5.6 Terra次旗舰Claude Sonnet 5次旗舰Claude Sonnet 4.6Gemini 3.1 Pro Preview
高深度推理与复杂任务
GPT-5.6 Sol旗舰GPT-5.5 InstantClaude Opus 4.8旗舰Claude Opus 4.7次旗舰Claude Opus 4.6次旗舰
很高旗舰创作与顶级任务
Claude Fable 5旗舰

完整定位与选型见 模型说明。

不同模型,积分消耗不同

高性价比模型适合日常高频任务,积分消耗低;旗舰与次旗舰模型适合重要决策与复杂分析,积分消耗相对更高。

原则:先选够用且更省积分的模型,只有任务确实需要更强能力时再升级。

查看模型说明 →

相关链接

  • 模型说明
  • 服务条款
  • 服务特定条款

Help & Docs

Credits & Billing

Credit rules, reservation, and relative cost levels.

Version
0.3.0
Effective date
2026-08-10
Last updated
2026-09-01

Credit basics

¥1 = 500 credits (1 credit = ¥0.002).

Credits come in two types — purchased and gifted — shown separately but both count as "available credits." Draw order is managed by the server; you do not need to choose manually.

What is credit reservation

When you send a request, the system reserves the maximum credits that request might use. After the request completes, actual usage is settled, and unused reservation is returned automatically (reservation return).

Reserved credits ≠ final charge. The amount reserved is a ceiling; actual deductions are usually lower.

What is model usage

Each successfully completed request creates a model usage record, settled by actual usage. Failed, cancelled, or empty-output requests release the reservation with no charge.

Minimum charge per request is 1 credit.

Credit cost levels

This page does not list precise per-model rates. Cost is shown as relative credit levels; final billing follows actual server-side usage.

LowGood for high-frequency daily use
OHLAC 1.0OHLAC 1.5GPT-5.6 LunaGreat valueGemini 3.5 FlashGreat valueGemini 3.7 FlashGreat valueGrok 4.5Great valueGrok 4.6Flagship
MediumBalanced tasks, moderate cost
OHLAC ProGPT-5.6 TerraNear-flagshipClaude Sonnet 5Near-flagshipClaude Sonnet 4.6Gemini 3.1 Pro Preview
HighDeep reasoning and complex work
GPT-5.6 SolFlagshipGPT-5.5 InstantClaude Opus 4.8FlagshipClaude Opus 4.7Near-flagshipClaude Opus 4.6Near-flagship
Very highFlagship creation and top-tier tasks
Claude Fable 5Flagship

Full positioning and selection tips: Models.

Credits vary by model

Great-value models suit everyday high-frequency tasks at low cost. Flagship and near-flagship models suit important decisions and complex analysis at higher cost.

Principle: start with the most affordable model that fits, and upgrade only when needed.

View model guide →

Related

  • Models
  • Terms of Service
  • Service-Specific Terms

Help & Docs

クレジットと課金

クレジットルール、予約、相対的な消費レベル。

バージョン
0.3.0
施行日
2026年8月10日
最終更新
2026年9月1日

Credit basics

¥1 = 500 credits (1 credit = ¥0.002).

Credits come in two types — purchased and gifted — shown separately but both count as "available credits." Draw order is managed by the server; you do not need to choose manually.

What is credit reservation

When you send a request, the system reserves the maximum credits that request might use. After the request completes, actual usage is settled, and unused reservation is returned automatically (reservation return).

Reserved credits ≠ final charge. The amount reserved is a ceiling; actual deductions are usually lower.

What is model usage

Each successfully completed request creates a model usage record, settled by actual usage. Failed, cancelled, or empty-output requests release the reservation with no charge.

Minimum charge per request is 1 credit.

Credit cost levels

This page does not list precise per-model rates. Cost is shown as relative credit levels; final billing follows actual server-side usage.

LowGood for high-frequency daily use
OHLAC 1.0OHLAC 1.5GPT-5.6 LunaGreat valueGemini 3.5 FlashGreat valueGemini 3.7 FlashGreat valueGrok 4.5Great valueGrok 4.6Flagship
MediumBalanced tasks, moderate cost
OHLAC ProGPT-5.6 TerraNear-flagshipClaude Sonnet 5Near-flagshipClaude Sonnet 4.6Gemini 3.1 Pro Preview
HighDeep reasoning and complex work
GPT-5.6 SolFlagshipGPT-5.5 InstantClaude Opus 4.8FlagshipClaude Opus 4.7Near-flagshipClaude Opus 4.6Near-flagship
Very highFlagship creation and top-tier tasks
Claude Fable 5Flagship

Full positioning and selection tips: Models.

Credits vary by model

Great-value models suit everyday high-frequency tasks at low cost. Flagship and near-flagship models suit important decisions and complex analysis at higher cost.

Principle: start with the most affordable model that fits, and upgrade only when needed.

View model guide →

Related

  • Models
  • Terms of Service
  • Service-Specific Terms

Help & Docs

Créditos y facturación

Reglas de créditos, reserva y niveles relativos de consumo.

Versión
0.3.0
Fecha de vigencia
10/8/2026
Última actualización
1/9/2026

Credit basics

¥1 = 500 credits (1 credit = ¥0.002).

Credits come in two types — purchased and gifted — shown separately but both count as "available credits." Draw order is managed by the server; you do not need to choose manually.

What is credit reservation

When you send a request, the system reserves the maximum credits that request might use. After the request completes, actual usage is settled, and unused reservation is returned automatically (reservation return).

Reserved credits ≠ final charge. The amount reserved is a ceiling; actual deductions are usually lower.

What is model usage

Each successfully completed request creates a model usage record, settled by actual usage. Failed, cancelled, or empty-output requests release the reservation with no charge.

Minimum charge per request is 1 credit.

Credit cost levels

This page does not list precise per-model rates. Cost is shown as relative credit levels; final billing follows actual server-side usage.

LowGood for high-frequency daily use
OHLAC 1.0OHLAC 1.5GPT-5.6 LunaGreat valueGemini 3.5 FlashGreat valueGemini 3.7 FlashGreat valueGrok 4.5Great valueGrok 4.6Flagship
MediumBalanced tasks, moderate cost
OHLAC ProGPT-5.6 TerraNear-flagshipClaude Sonnet 5Near-flagshipClaude Sonnet 4.6Gemini 3.1 Pro Preview
HighDeep reasoning and complex work
GPT-5.6 SolFlagshipGPT-5.5 InstantClaude Opus 4.8FlagshipClaude Opus 4.7Near-flagshipClaude Opus 4.6Near-flagship
Very highFlagship creation and top-tier tasks
Claude Fable 5Flagship

Full positioning and selection tips: Models.

Credits vary by model

Great-value models suit everyday high-frequency tasks at low cost. Flagship and near-flagship models suit important decisions and complex analysis at higher cost.

Principle: start with the most affordable model that fits, and upgrade only when needed.

View model guide →

Related

  • Models
  • Terms of Service
  • Service-Specific Terms
© 2026 OHLAC
说明中心 模型说明积分与扣费