26 models from 6 providers — Anthropic, OpenAI, DeepSeek, Google, Z.AI and xAI — all running on a single api.chotaikhoan.com key. Buy a quota package, get charged transparently by a public formula, no international card needed.
Each model's charging rate is printed right on its card. No hidden fees, no silent charges.
The key is yours
Your key is only ever shown to you. Lost it? Restore it with your original redeem code.
System errors are free
Requests that fail on our side are clearly marked as not charged in your history.
Real human support
Stuck at any step? Message us on Telegram @chotaikhoandotcom — we help you set everything up, step by step.
Get started in 3 steps
1
Buy a redeem code
Pick a package on the Packages page and pay — your redeem code appears right on the page.
2
Redeem it for a key
Open the Redeem page and enter your code — your key appears instantly.
3
Paste it into your app
Follow the Guide page: 3 steps for each app, with ready-made sample configs.
Quota is charged by the actual usage of each request, based on its input/output tokens and the model you choose, at the rates on the Models & pricing page. Going through Claude Code consumes quota faster than chat and Codex, because it is a specialised coding route with higher processing costs. Image generation is charged per image. Failed requests are not charged, and you can check the exact quota of every request on the Check key page.See detailed conversion rates by calling path
Detailed charging: quota charged = input tokens × input price + output tokens × output price + cache tokens × cache price of the model — each model's rates are printed on its card on the Models & pricing page. Conversion rate by calling route: via Claude Code (the /v1/messages route) ×4.1; via chat and Codex (the /v1/chat/completions, /v1/responses routes) ×2.25. Image generation is charged a fixed amount per image.
Automatic delivery 24/7Your code arrives as soon as payment lands — no waiting
🛡
100% refund within 7 daysIf the key fails because of us
💬
Support on TelegramReal humans help when you need it — @chotaikhoandotcom
Quota is charged by the actual usage of each request, based on its input/output tokens and the model you choose, at the rates on the Models & pricing page. Going through Claude Code consumes quota faster than chat and Codex, because it is a specialised coding route with higher processing costs. Image generation is charged per image. Failed requests are not charged, and you can check the exact quota of every request on the Check key page.See detailed conversion rates by calling path
Detailed charging: quota charged = input tokens × input price + output tokens × output price + cache tokens × cache price of the model — each model's rates are printed on its card on the Models & pricing page. Conversion rate by calling route: via Claude Code (the /v1/messages route) ×4.1; via chat and Codex (the /v1/chat/completions, /v1/responses routes) ×2.25. Image generation is charged a fixed amount per image.
How to buy
1
Pick a quantity and a payment method right on the package card
Adjust the quantity if you want several packages, open the payment method box and choose USDT, VPBank or MoMo — the payment box opens right away and bank/USDT orders stay valid for 24 hours, MoMo for 30 minutes. For international buyers, USDT (TRC20) is the smoothest option.
2
Pay the exact amount shown
For bank transfer, scan the QR and keep the transfer note exactly as shown — the system recognises your order by it. For USDT, send the exact unique amount of your order so the system can match it automatically.
3
Your code appears right on the page
As soon as payment lands (usually seconds to a few minutes), your redeem code shows up — redeem it for a key and start working.
Prefer a hand-held purchase? Message us on Telegram @chotaikhoandotcom — we also sell manually via USDT / MoMo / bank transfer.
All 26 models from 6 providers on a single key. Charge rates are printed on each card on the Models & pricing page.
Models & pricing
All 26 models from 6 providers on a single api.chotaikhoan.com key. Tap a provider name to filter. Prices shown are international reference prices (USD / 1M tokens); the struck-through figures are the international list prices. At api.chotaikhoan.com you do not pay per request: you buy a quota package, and quota is charged at each model's own rates: input tokens × input price + output tokens × output price + cache tokens × cache price — the rates are printed on every card below; each calling route has its own conversion rate based on that route's processing cost — see the box at the bottom of this page for details.
Showing 26 / 26 models
Anthropic7 models
claude-opus-5-5
Anthropic
Claude mạnh nhất hiện có: suy luận sâu, dùng công cụ, phân tích ảnh. Để dành cho việc khó nhất.
IN$2.99$4
OUT$15.96$20
CACHE$0.0792$0.2
CHARGE RATE / 1Min 1 · out 1
-21% vs list price
claude-opus-5
Anthropic
Bản Opus cân bằng: vẫn rất mạnh ở suy luận và code, đơn giá mềm hơn hẳn bản 5-5.
IN$0.992$15
OUT$5.96$75
CACHE$0.0792$1.5
CHARGE RATE / 1Min 1 · out 1
-92% vs list price
claude-sonnet-5
Anthropic
Con ngựa thồ của dân code: nhanh, khôn, giá vừa — lựa chọn mặc định cho Claude Code và Cline.
IN$0.35$6
OUT$1.65$30
CACHE$0.035$0.9
CHARGE RATE / 1Min 1 · out 1
-94% vs list price
claude-opus-4-8
Anthropic
Opus thế hệ trước: suy luận và phân tích ảnh vẫn tốt, giá đã mềm đi nhiều.
IN$0.792$15
OUT$2.96$75
CACHE$0.0792$1.5
CHARGE RATE / 1Min 1 · out 1
-96% vs list price
claude-fable-5-1
Anthropic
Dòng Fable bản nâng cấp: viết lách, kể chuyện và suy luận bằng ngôn ngữ rất mượt.
IN$1.53$10
OUT$7.64$50
CACHE$0.0528$0.25
CHARGE RATE / 1Min 1 · out 1
-85% vs list price
claude-fable-5
Anthropic
Dòng Fable: mạnh về ngôn ngữ, viết và phân tích ảnh.
IN$1.53$10
OUT$9.64$50
CACHE$0.0528$1
CHARGE RATE / 1Min 1 · out 1
-81% vs list price
claude-haiku-4-5-20251001
Anthropic
Nhanh và rẻ nhất nhà Claude: hợp việc nhẹ, số lượng lớn, cần tốc độ.
IN$0.08$3
OUT$0.35$15
CACHE$0.008$0.3
CHARGE RATE / 1Min 1 · out 1
-98% vs list price
OpenAI11 models
gpt-6-astra
OpenAI
Đầu tàu của OpenAI: suy luận sâu và gọi công cụ cho tác vụ phức tạp nhất.
IN$2.2$18
OUT$11.2$90
CACHE$0.22$1.8
CHARGE RATE / 1Min 1 · out 1
-88% vs list price
gpt-5.6-sol
OpenAI
GPT bản mạnh cho công cụ và code, đơn giá dễ chịu hơn Astra.
IN$0.792$4
OUT$2.96$20
CACHE$0.0264$0.55
CHARGE RATE / 1Min 1 · out 1
-84% vs list price
gpt-5.6-terra
OpenAI
GPT đa dụng: viết, tóm tắt, việc văn phòng — nhanh và rẻ.
IN$0.2$2
OUT$1$12
CACHE$0.02$0.3
CHARGE RATE / 1Min 1 · out 1
-91% vs list price
gpt-5.5
OpenAI
GPT cân bằng: gọi công cụ, phân tích ảnh — hợp dùng hằng ngày trong Codex và Cursor.
IN$0.24$5
OUT$1.44$30
CACHE$0.024$0.5
CHARGE RATE / 1Min 1 · out 1
-95% vs list price
gpt-5.4
OpenAI
GPT giá mềm mà vẫn đọc được ảnh và gọi công cụ.
IN$0.08$2.5
OUT$0.35$15
CACHE$0.008$0.25
CHARGE RATE / 1Min 1 · out 1
-98% vs list price
gpt-5.4-mini
OpenAI
Bản mini: nhanh, rẻ — cho tác vụ lặp lại với số lượng lớn.
IN$0.08$0.75
OUT$0.35$4.5
CACHE$0.008$0.075
CHARGE RATE / 1Min 1 · out 1
-92% vs list price
gpt-5.6-luna
OpenAI
Rẻ nhất nhà GPT: việc nhẹ, thử nghiệm, chạy nền.
IN$0.08$0.2
OUT$0.35$1.2
CACHE$0.008$0.15
CHARGE RATE / 1Min 1 · out 1
-69% vs list price
codex-auto-review
OpenAI
Chuyên rà soát code: soi bản thay đổi, tìm lỗi và rủi ro trước khi gộp code.
IN$0.25$3
OUT$1$15
CACHE$0.025
CHARGE RATE / 1Min 1 · out 1
-93% vs list price
gpt-oss-120b
OpenAI
Model mã nguồn mở 120B: tự do, rẻ, hợp thử nghiệm.
IN$0.08
OUT$0.35
CACHE$0.008
CHARGE RATE / 1Min 1 · out 1
Reference price
gpt-image-2
OpenAI
Tạo ảnh từ chữ: banner, ảnh minh họa, ảnh sản phẩm.
PER REQUEST$0.024$0.5
QUOTA CHARGE500,000 tokens/image
-95% vs list price
gpt-image-2.5
OpenAI
Tạo ảnh bản mới: chi tiết hơn, bám sát yêu cầu hơn.
PER REQUEST$0.024$0.5
QUOTA CHARGE500,000 tokens/image
-95% vs list price
DeepSeek3 models
deepseek-v4-pro
DeepSeek
DeepSeek V4 bản mạnh: code và suy luận dài rất khá, giá kiểu Trung Quốc.
IN$1.45
OUT$4.38
CHARGE RATE / 1Min 1 · out 1
Reference price
deepseek-v4-flash
DeepSeek
Bản nhanh của DeepSeek V4: rẻ, hợp chạy hàng loạt và code thường ngày.
IN$0.45
OUT$1.38
CHARGE RATE / 1Min 1 · out 1
Reference price
deepseek-v4.1-flash-1017
DeepSeek
Rẻ gần như cho không: hợp thử nghiệm và tác vụ siêu nhẹ.
IN$0.0003
OUT$0.0009
CHARGE RATE / 1Min 1 · out 1
Reference price
Google3 models
gemini-3.8-flash
Google
Gemini bản nhanh: phản hồi tức thì, đọc ảnh, gọi công cụ.
IN$0.08$1.5
OUT$0.35$7.5
CACHE$0.008$0.15
CHARGE RATE / 1Min 1 · out 1
-95% vs list price
gemini-3.7-flash
Google
Gemini bản nhanh thế hệ trước: giá mềm, đủ dùng hằng ngày.
IN$0.08$1.5
OUT$0.35$7.5
CACHE$0.008$0.15
CHARGE RATE / 1Min 1 · out 1
-95% vs list price
gemini-3.1-flash-lite-image
Google
Tạo ảnh bản nhẹ: nhanh và rẻ khi cần số lượng nhiều.
PER REQUEST$0.012$0.25
QUOTA CHARGE500,000 tokens/image
-95% vs list price
Z.AI1 models
glm-5.3
Z.AI
Đầu tàu của Z.AI: mạnh về văn bản song ngữ và xử lý ngữ cảnh rất dài.
IN$3.15
OUT$9.9
CHARGE RATE / 1Min 1 · out 1
Reference price
xAI1 models
grok-4.70
xAI
Grok của xAI: phản hồi nhanh, gọi công cụ, trả lời thẳng và gọn.
IN$0.08$1.25
OUT$0.35$2.5
CACHE$0.008$0.2
CHARGE RATE / 1Min 1 · out 1
-89% vs list price
International reference prices are only there for comparison. Your actual money is the quota in your package: input tokens × input price + output tokens × output price + cache tokens × cache price of the model — conversion rate by calling route: via Claude Code (/v1/messages) ×4.1, and via chat and Codex (/v1/chat/completions, /v1/responses) ×2.25. Every model shares one key — switch models right inside your app, no extra purchase needed.
Reference prices last updated: 02/11/2026.
Check key
Paste your key to see your quota, validity and per-model usage detail.
Your key is saved in this browser only — reopen this page (even after closing the tab or shutting down) and your balance is right there, no need to paste again. Paste a different key and press Check to replace the saved one. On a shared computer, press when you are done.
1
Paste your key
The key you received when redeeming a code — it starts with sk-ctk-.
2
Press Check
The system looks it up instantly — no login needed.
3
See the result
Remaining quota, validity, daily usage and per-model detail.
4
The key is yours alone
Results only appear with the exact key. Never share your key with anyone.
Small note: the tokens here are this service's internal balance, not the model's native tokens — how much each request charges depends on the model and the context length. Each request is charged at the model's own rates: input tokens × input price + output tokens × output price + cache tokens × cache price — each model's rates are printed on its card on the Models & pricing page. The Claude Code route (/v1/messages) converts at ×4.1, because every Claude Code request carries the full conversation history, a very long system prompt and tool descriptions. The chat and Codex routes (/v1/chat/completions, /v1/responses) convert at ×2.25. Image generation (gpt-image-2, gpt-image-2.5): 500,000 tokens per image. Your key is a retail package split by the shop, so it expires exactly as the package you bought — see the "Expires" line in the result above.
Daily usage
Per-model detail (last 24 hours)
One row per model: how many requests you made and exactly how much quota was charged at the rates on the model card.
Recent request history
TIME
MODEL
STATUS
INPUT TOKENS
OUTPUT TOKENS
QUOTA CHARGED
Ready-made config for this key
Your key is already filled into the samples below — copy, paste into your app, and go.
Claude Code
Cursor / Cline (OpenAI Compatible)
Codex CLI
Why does Claude Code burn tokens so much faster than regular chat? It is not a trick: with every request, Claude Code sends along the entire conversation history, a very long system prompt and the descriptions of every tool it is allowed to use — so the input tokens of a single request are already many times those of a normal chat message. Add the background requests (context summaries, session titles...) and the token count climbs very fast even if you only type a few commands. That is why the Claude Code route charges: (input tokens × input price + output tokens × output price + cache tokens × cache price) × 4.1. The chat and Codex routes use the same calculation, converted at ×2.25.
Order lookup & payment receipt
Enter the CTK… order code you received when buying to see the order status, get the QR code again if you have not paid yet, or print a confirmation receipt after paying.
Setup guide
Just 3 steps, no technical skills needed. Stuck anywhere? Message us on Telegram @chotaikhoandotcom — we will walk you through it. Our Telegram bot also has the full setup guide in English.
1
Buy a code
Pick a package on the Packages page and pay — you receive a redeem code.
2
Redeem it for a key
Open the Redeem page and enter the code — you get a key starting with sk-ctk-.
3
Paste the key into your app
Pick your app below, copy its config snippet, paste your key in — done.
Every app has its configuration snippet right on its card — copy it, replace YOUR_API_KEY with your key, and you are running. Every app below has a full step-by-step article in English:
Claude Code
Full article: installation, environment variables and the most common error fixes.
The three most-used setups, with your saved key filled in automatically when you have one:
Claude Code
Cursor / Cline (OpenAI Compatible)
Codex CLI
The magic check sentence: in your app, type exactly this: "Reply with exactly one word: running". If the app answers "running", your setup is correct.
Hit an error? Check this table first
Invalid key error
Usually a missing character or an extra space at the start/end when copying. Copy the key again — never type it by hand.
Model not found
The model name is typed wrong. Copy the exact name from the Models & pricing page — there is a Copy button in each model's Details box.
Out of quota
Your key has used up its package. Redeem another code on the Redeem page — quota stacks onto your old key and validity follows the new package.
Empty replies
Your app has max_tokens set too low. Raise max_tokens to 500 or more.
System busy
A temporary error on our side — that request was not charged. Wait a few minutes and try again.
Still stuck
Message us on Telegram @chotaikhoandotcom with the time of the error, the model name and the error line — we will take it from there.
The technical bits (only needed when your app asks): OpenAI address: https://api.chotaikhoan.com/v1 — Anthropic address: https://api.chotaikhoan.com. One key works for every app and every model.
No key yet?
The Trial package is only 19.000đ — enough to install and test every app yourself before buying a bigger package.
Enter the redeem code the shop sent you (format CTK-XXXX-XXXX-XXXX). Unused codes stay valid for 30 days.
The quota is added to your old key and the validity follows the new package. Your old key stays exactly the same — no need to set up your apps again.
Lost your key? Re-enter the exact original redeem code of that key in the box below — the original code is your proof of ownership, and the system will show your key again.
No code yet? See Packages and pay right on the page, or message us on Telegram @chotaikhoandotcom to buy.
claude-opus-5-5
Anthropic
Claude mạnh nhất hiện có: suy luận sâu, dùng công cụ, phân tích ảnh. Để dành cho việc khó nhất.
IN / 1M$2.99$4
OUT / 1M$15.96$20
CACHE / 1M$0.0792$0.2
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-opus-5
Anthropic
Bản Opus cân bằng: vẫn rất mạnh ở suy luận và code, đơn giá mềm hơn hẳn bản 5-5.
IN / 1M$0.992$15
OUT / 1M$5.96$75
CACHE / 1M$0.0792$1.5
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-sonnet-5
Anthropic
Con ngựa thồ của dân code: nhanh, khôn, giá vừa — lựa chọn mặc định cho Claude Code và Cline.
IN / 1M$0.35$6
OUT / 1M$1.65$30
CACHE / 1M$0.035$0.9
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-opus-4-8
Anthropic
Opus thế hệ trước: suy luận và phân tích ảnh vẫn tốt, giá đã mềm đi nhiều.
IN / 1M$0.792$15
OUT / 1M$2.96$75
CACHE / 1M$0.0792$1.5
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-fable-5-1
Anthropic
Dòng Fable bản nâng cấp: viết lách, kể chuyện và suy luận bằng ngôn ngữ rất mượt.
IN / 1M$1.53$10
OUT / 1M$7.64$50
CACHE / 1M$0.0528$0.25
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-fable-5
Anthropic
Dòng Fable: mạnh về ngôn ngữ, viết và phân tích ảnh.
IN / 1M$1.53$10
OUT / 1M$9.64$50
CACHE / 1M$0.0528$1
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
claude-haiku-4-5-20251001
Anthropic
Nhanh và rẻ nhất nhà Claude: hợp việc nhẹ, số lượng lớn, cần tốc độ.
IN / 1M$0.08$3
OUT / 1M$0.35$15
CACHE / 1M$0.008$0.3
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-6-astra
OpenAI
Đầu tàu của OpenAI: suy luận sâu và gọi công cụ cho tác vụ phức tạp nhất.
IN / 1M$2.2$18
OUT / 1M$11.2$90
CACHE / 1M$0.22$1.8
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.6-sol
OpenAI
GPT bản mạnh cho công cụ và code, đơn giá dễ chịu hơn Astra.
IN / 1M$0.792$4
OUT / 1M$2.96$20
CACHE / 1M$0.0264$0.55
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.6-terra
OpenAI
GPT đa dụng: viết, tóm tắt, việc văn phòng — nhanh và rẻ.
IN / 1M$0.2$2
OUT / 1M$1$12
CACHE / 1M$0.02$0.3
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.5
OpenAI
GPT cân bằng: gọi công cụ, phân tích ảnh — hợp dùng hằng ngày trong Codex và Cursor.
IN / 1M$0.24$5
OUT / 1M$1.44$30
CACHE / 1M$0.024$0.5
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.4
OpenAI
GPT giá mềm mà vẫn đọc được ảnh và gọi công cụ.
IN / 1M$0.08$2.5
OUT / 1M$0.35$15
CACHE / 1M$0.008$0.25
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.4-mini
OpenAI
Bản mini: nhanh, rẻ — cho tác vụ lặp lại với số lượng lớn.
IN / 1M$0.08$0.75
OUT / 1M$0.35$4.5
CACHE / 1M$0.008$0.075
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-5.6-luna
OpenAI
Rẻ nhất nhà GPT: việc nhẹ, thử nghiệm, chạy nền.
IN / 1M$0.08$0.2
OUT / 1M$0.35$1.2
CACHE / 1M$0.008$0.15
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
codex-auto-review
OpenAI
Chuyên rà soát code: soi bản thay đổi, tìm lỗi và rủi ro trước khi gộp code.
IN / 1M$0.25$3
OUT / 1M$1$15
CACHE / 1M$0.025
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-oss-120b
OpenAI
Model mã nguồn mở 120B: tự do, rẻ, hợp thử nghiệm.
IN / 1M$0.08
OUT / 1M$0.35
CACHE / 1M$0.008
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gpt-image-2
OpenAI
Tạo ảnh từ chữ: banner, ảnh minh họa, ảnh sản phẩm.
PER REQUEST$0.024$0.5
Charge: 500,000 tokens / image — image models charge a fixed amount per image, not by input/output tokens.
Paste this model name into your app config — see the Guide page.
gpt-image-2.5
OpenAI
Tạo ảnh bản mới: chi tiết hơn, bám sát yêu cầu hơn.
PER REQUEST$0.024$0.5
Charge: 500,000 tokens / image — image models charge a fixed amount per image, not by input/output tokens.
Paste this model name into your app config — see the Guide page.
deepseek-v4-pro
DeepSeek
DeepSeek V4 bản mạnh: code và suy luận dài rất khá, giá kiểu Trung Quốc.
IN / 1M$1.45
OUT / 1M$4.38
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
deepseek-v4-flash
DeepSeek
Bản nhanh của DeepSeek V4: rẻ, hợp chạy hàng loạt và code thường ngày.
IN / 1M$0.45
OUT / 1M$1.38
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
deepseek-v4.1-flash-1017
DeepSeek
Rẻ gần như cho không: hợp thử nghiệm và tác vụ siêu nhẹ.
IN / 1M$0.0003
OUT / 1M$0.0009
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gemini-3.8-flash
Google
Gemini bản nhanh: phản hồi tức thì, đọc ảnh, gọi công cụ.
IN / 1M$0.08$1.5
OUT / 1M$0.35$7.5
CACHE / 1M$0.008$0.15
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gemini-3.7-flash
Google
Gemini bản nhanh thế hệ trước: giá mềm, đủ dùng hằng ngày.
IN / 1M$0.08$1.5
OUT / 1M$0.35$7.5
CACHE / 1M$0.008$0.15
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
gemini-3.1-flash-lite-image
Google
Tạo ảnh bản nhẹ: nhanh và rẻ khi cần số lượng nhiều.
PER REQUEST$0.012$0.25
Charge: 500,000 tokens / image — image models charge a fixed amount per image, not by input/output tokens.
Paste this model name into your app config — see the Guide page.
glm-5.3
Z.AI
Đầu tàu của Z.AI: mạnh về văn bản song ngữ và xử lý ngữ cảnh rất dài.
IN / 1M$3.15
OUT / 1M$9.9
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
grok-4.70
xAI
Grok của xAI: phản hồi nhanh, gọi công cụ, trả lời thẳng và gọn.
IN / 1M$0.08$1.25
OUT / 1M$0.35$2.5
CACHE / 1M$0.008$0.2
Charge rates: in 1 · out 1 · cache 1 (quota tokens per token used). Example: a request with 10,000 input tokens + 2,000 output tokens → charged 12,000 tokens. Conversion by calling route — via Claude Code (the /v1/messages route): ×4.1; via chat or Codex (the /v1/chat/completions, /v1/responses routes): ×2.25.
Paste this model name into your app config — see the Guide page.
Payment method
Starter package
$0.72≈ 19.000đ · 10M tokens · valid 1 day
⏱Picking a method opens the payment box right away — no extra tap needed. Bank-transfer and USDT orders are valid for 24 hours, MoMo orders for 30 minutes, with a countdown in the payment box; if time runs out, just create a new order. Prices are in Vietnamese đồng (VND) — USD figures are for reference, and USDT orders are settled at the exact USD amount shown.
Pro package
$1.70≈ 45.000đ · 100M tokens · valid 2 days
⏱Picking a method opens the payment box right away — no extra tap needed. Bank-transfer and USDT orders are valid for 24 hours, MoMo orders for 30 minutes, with a countdown in the payment box; if time runs out, just create a new order. Prices are in Vietnamese đồng (VND) — USD figures are for reference, and USDT orders are settled at the exact USD amount shown.
Max package
$5.25≈ 139.000đ · 500M tokens · valid 7 days
⏱Picking a method opens the payment box right away — no extra tap needed. Bank-transfer and USDT orders are valid for 24 hours, MoMo orders for 30 minutes, with a countdown in the payment box; if time runs out, just create a new order. Prices are in Vietnamese đồng (VND) — USD figures are for reference, and USDT orders are settled at the exact USD amount shown.
Ultra package
$16.94≈ 449.000đ · 2B tokens · valid 30 days
⏱Picking a method opens the payment box right away — no extra tap needed. Bank-transfer and USDT orders are valid for 24 hours, MoMo orders for 30 minutes, with a countdown in the payment box; if time runs out, just create a new order. Prices are in Vietnamese đồng (VND) — USD figures are for reference, and USDT orders are settled at the exact USD amount shown.