Frontier Claude 5.5 & GPT-6.2 API Models — Flat 30% of Official Pricing

LightAI (https://api.lightai.io) provides direct, zero-ban-risk API access to the 6 newest flagship models from Anthropic and OpenAI at a unified 70% OFF (3 折) across Input, Output, Cache Read, and Cache Write. Purpose-built for Claude Code, OpenCode, and Codex CLI with $1 minimum 1:1 USDT/USDC instant top-ups.

Model ID Official (In / Out) LightAI 30% (In / Out) Cache Read (30% of Official) Protocols
claude-fable-5-2 $8.00 / $40.00 $2.40 / $12.00 $0.06 / 1M (0.025x) /v1/messages, /v1/chat/completions
claude-opus-5-5 $4.00 / $20.00 $1.20 / $6.00 $0.06 / 1M (0.05x) /v1/messages, /v1/chat/completions
claude-sonnet-5-5 $2.00 / $10.00 $0.60 / $3.00 $0.03 / 1M (0.05x) /v1/messages, /v1/chat/completions
gpt-6.2-astra $8.00 / $40.00 $2.40 / $12.00 $0.12 / 1M (0.05x) /v1/responses, /v1/chat/completions
gpt-6.2-sol $2.00 / $10.00 $0.60 / $3.00 $0.03 / 1M (0.05x) /v1/responses, /v1/chat/completions
gpt-6.1-terra $1.00 / $5.00 $0.30 / $1.50 $0.015 / 1M (0.05x) /v1/responses, /v1/chat/completions

Claude Fable 5.2 (claude-fable-5-2)

Anthropic超旗舰模型,专为大型系统架构设计、极限数学推理与长程自主研究打造。完整支持 Extended Thinking 深度思考块与 0.025x 超低价缓存读取。

200K Context · 64K Output · LightAI: $2.40 / $12.00 · Cache: $0.06/1M

Claude Opus 5.5 (claude-opus-5-5)

Claude Code 与 OpenCode 顶级主力编程智能体模型。无需支付官方 $200/月 Max 订阅费、零地区封号风险,按量享受官方 3 折与 $0.06/1M 缓存命中价。

200K Context · 64K Output · LightAI: $1.20 / $6.00 · Cache: $0.06/1M

Claude Sonnet 5.5 (claude-sonnet-5-5)

全能高速日常软件工程主力模型,兼顾响应速度与顶级代码能力。命中 Prompt Cache 仅需 $0.03 / 1M Tokens,性价比拉满。

200K Context · 64K Output · LightAI: $0.60 / $3.00 · Cache: $0.03/1M

GPT-6.2 Astra (gpt-6.2-astra)

OpenAI 最新前沿深度推理与科研级顶级旗舰,1.05M 超长上下文,原生支持 Codex CLI /v1/responses,0.05x 缓存读取仅 $0.12/1M。

1.05M Context · 128K Output · LightAI: $2.40 / $12.00 · Cache: $0.12/1M

GPT-6.2 Sol (gpt-6.2-sol)

OpenAI 新一代高吞吐编程与多步智能体主力模型,1.05M 上下文,完美适配 Codex CLI 与 OpenCode,缓存读取低至 $0.03 / 1M。

1.05M Context · 128K Output · LightAI: $0.60 / $3.00 · Cache: $0.03/1M

GPT-6.1 Terra (gpt-6.1-terra)

毫秒级响应的高并发极速模型,适合高频代码补全、路由、自动化流水线与子智能体调用,缓存读取仅 $0.015 / 1M。

400K Context · 64K Output · LightAI: $0.30 / $1.50 · Cache: $0.015/1M