💰 Models & Pricing
Old Dog AI is committed to providing a high‑value, fully transparent AI API service. We do not use expensive monthly subscriptions—instead, we adopt a pay‑as‑you‑go model. You only pay for what you actually use—no waste.
🪙 Billing principles & balance
- Billing unit: We use standard USD credits internally. Charges align closely with the official token accounting logic.
- Check balance: After signing in to the Old Dog AI Console, you can view available balance and spending history on the home dashboard, or via the avatar menu in the top-right corner.
- Real‑time rates / multipliers: Upstream official pricing changes frequently. Our pricing adjusts dynamically as well. Please refer to the real‑time multiplier shown on the Console’s “Pricing” page.
🤖 Recommended models (Updated for 2026)
We aggregate the world’s leading LLMs. In your apps or code, make sure to enter the exact “Model Name” shown below, otherwise you may get “model not found” errors.
🟣 Anthropic series (developers’ default choice)
| Recommended scenario | Exact model name (use this) | Notes |
|---|---|---|
| Top frontier flagship | claude-opus-4-6 | Latest top-tier intelligence model with strong agentic capabilities, suitable for complex multi-repo architecture tasks. |
| Daily dev workhorse (recommended) | claude-sonnet-4-6 | Excellent for code and text generation; big leap in Computer Use; widely regarded as the best developer model. |
🟢 OpenAI & Codex series (deep reasoning & code completion)
| Recommended scenario | Exact model name (use this) | Notes |
|---|---|---|
| General flagship | gpt-5.2 | OpenAI’s all‑round flagship model with stable performance and strong multimodal experience. |
| Top deep reasoning | o3-pro | Great for very hard math/architecture problems—thinks deeper and aims to get it right in one pass. |
| Dedicated code completion | gpt-5.3-codex | Tailored for IDEs (Cursor, VS Code) and complex completions; very fast response. |
🔵 Google series (latest multimodal powerhouse)
| Recommended scenario | Exact model name (use this) | Notes |
|---|---|---|
| Latest general flagship | gemini-3.2-pro | Google’s latest 3.2‑architecture model, strong across benchmarks, competitive with the best in the industry. |
| Ultra‑long context | gemini-2.5-pro | Mature and stable with very long context windows—ideal for reading books or analyzing large codebases. |
🟠 xAI (Grok) series (real‑time info & uncensored style)
| Recommended scenario | Exact model name (use this) | Notes |
|---|---|---|
| All‑round humorous flagship | grok-3-pro | xAI’s latest flagship; strong at witty, geeky conversation and real‑time internet trends. |
🔴 Kimi (Moonshot) series (Chinese language leader)
| Recommended scenario | Exact model name (use this) | Notes |
|---|---|---|
| Long Chinese documents | kimi-latest | Excellent Chinese fluency and long‑text memory; great for novels, financial reports, and meeting transcripts. |
| Chinese coding assistance | kimi-for-coding | Optimized for Chinese developers; explanations and comments match common local preferences. |
💡 Tip: Beyond the curated set above, we also support dozens of popular/open models. For the full list, check Models in the console.
💳 Top up & activation
When your trial credits are used up, you can top up via the following. Credits are added instantly after payment succeeds:
- Alipay: QR-code payment with automatic FX conversion—best for users in Mainland China.
- International card (Stripe): Visa/Mastercard/Apple Pay, etc.—for overseas users and global teams (redirects to Stripe secure checkout).
- Crypto: USDT (TRC20, etc.) for privacy‑preserving transfers.