Balance tokens charged per real token. 1 balance token = 1 Claude Sonnet 5.5 input token. Dollar prices are per 1M tokens, input / output.
Claude
Anthropic’s models for code, writing and agents. Haiku is fast and cheap, Sonnet is the everyday workhorse, Opus and Fable take the hardest tasks. The natural pick for Claude Code.
Claude API| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Claude Fable 5.1 | ×5 / ×25 | $10 / $50 | $1 / $5 | |
| Claude Fable 5 | ×5 / ×25 | $10 / $50 | $1 / $5 | |
| Claude Opus 5.5 | ×2 / ×10 | $4 / $20 | $0.4 / $2 | |
| Claude Opus 5 | ×2.5 / ×12.5 | $5 / $25 | $0.5 / $2.5 | |
| Claude Opus 4.8 | ×2.5 / ×12.5 | $5 / $25 | $0.5 / $2.5 | |
| Claude Opus 4.7 | ×2.5 / ×12.5 | $5 / $25 | $0.5 / $2.5 | |
| Claude Opus 4.6 | ×2.5 / ×12.5 | $5 / $25 | $0.5 / $2.5 | |
| Claude Opus 4.5 | ×2.5 / ×12.5 | $5 / $25 | $0.5 / $2.5 | |
| Claude Sonnet 5.5 | ×1 / ×5 | $2 / $10 | $0.2 / $1 | |
| Claude Sonnet 5 | ×1 / ×5 | $2 / $10 | $0.2 / $1 | |
| Claude Sonnet 4.6 | ×1.5 / ×7.5 | $3 / $15 | $0.3 / $1.5 | |
| Claude Sonnet 4.5 | ×1.5 / ×7.5 | $3 / $15 | $0.3 / $1.5 | |
| Claude Haiku 5.5 | ×0.05 / ×0.25 | $0.1 / $0.5 | $0.01 / $0.05 | |
| Claude Haiku 4.5 | ×0.5 / ×2.5 | $1 / $5 | $0.1 / $0.5 |
GPT
OpenAI’s models, from tiny Nano and Luna for bulk requests to Astra for heavy reasoning. They work with Codex CLI, the OpenAI SDK and any bot.
ChatGPT API| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| GPT-6.1 Sol | ×1 / ×5 | $2 / $10 | $0.2 / $1 | |
| GPT-6 Astra | ×5 / ×25 | $10 / $50 | $1 / $5 | |
| GPT-6.1 Terra | ×0.5 / ×2.5 | $1 / $5 | $0.1 / $0.5 | |
| GPT-6 Luna | ×0.05 / ×0.25 | $0.1 / $0.5 | $0.01 / $0.05 | |
| GPT-5.4 Mini | ×0.38 / ×2.25 | $0.75 / $4.5 | $0.075 / $0.45 | |
| GPT-5.4 Nano | ×0.1 / ×0.63 | $0.2 / $1.25 | $0.02 / $0.125 |
Gemini
Google’s long-context models: Flash for fast, cheap answers, Pro and Argon for large documents and codebases.
Gemini API| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Gemini 4 Argon | ×1 / ×6 | $2 / $12 | $0.2 / $1.2 | |
| Gemini 3.1 Pro | ×1 / ×6 | $2 / $12 | $0.2 / $1.2 | |
| Gemini 3 Pro | ×1 / ×6 | $2 / $12 | $0.2 / $1.2 | |
| Gemini 2.5 Pro | ×0.63 / ×5 | $1.25 / $10 | $0.125 / $1 | |
| Gemini 3 Flash | ×0.25 / ×1.5 | $0.5 / $3 | $0.05 / $0.3 | |
| Gemini 3.6 Flash | ×0.25 / ×1.5 | $0.5 / $3 | $0.05 / $0.3 | |
| Gemini 3.5 Flash | ×0.25 / ×1.5 | $0.5 / $3 | $0.05 / $0.3 | |
| Gemini 2.5 Flash | ×0.15 / ×1.25 | $0.3 / $2.5 | $0.03 / $0.25 | |
| Gemini 2.5 Flash Lite | ×0.05 / ×0.2 | $0.1 / $0.4 | $0.01 / $0.04 |
Grok
xAI’s models: strong reasoning, fresh knowledge and a dedicated Grok Build for programming.
Grok API| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Grok 4.7 | ×1 / ×3 | $2 / $6 | $0.2 / $0.6 | |
| Grok 4.6 | ×1 / ×3 | $2 / $6 | $0.2 / $0.6 | |
| Grok 4.5 | ×1 / ×3 | $2 / $6 | $0.2 / $0.6 | |
| Grok 4.3 | ×0.63 / ×1.25 | $1.25 / $2.5 | $0.125 / $0.25 | |
| Grok Build 0.1 | ×0.5 / ×1 | $1 / $2 | $0.1 / $0.2 |
DeepSeek
The best-value models for code and reasoning. Flash for bulk work, V4 Pro for the hard parts.
DeepSeek API| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| DeepSeek Flash | ×0.15 / ×0.6 | $0.3 / $1.2 | $0.03 / $0.12 | |
| DeepSeek V4 Pro | ×0.66 / ×1.98 | $1.32 / $3.96 | $0.132 / $0.396 |
Kimi
Moonshot AI’s models for long tasks and agents; the Thinking version reasons before it answers.
| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Kimi K2 | ×0.3 / ×1.25 | $0.6 / $2.5 | $0.06 / $0.25 | |
| Kimi K2 Thinking | ×0.3 / ×1.25 | $0.6 / $2.5 | $0.06 / $0.25 |
Qwen
Alibaba Cloud’s models: Qwen3 Max for general work, Qwen3 Coder for programming.
| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Qwen3 Max | ×0.6 / ×3 | $1.2 / $6 | $0.12 / $0.6 | |
| Qwen3 Coder | ×0.5 / ×2.5 | $1 / $5 | $0.1 / $0.5 |
Mistral
European models by Mistral AI: Large and Medium for text, Codestral for code completion.
| Model | API ID | Balance rate | Official | KLINCH |
|---|---|---|---|---|
| Mistral Large | ×1 / ×3 | $2 / $6 | $0.2 / $0.6 | |
| Mistral Medium | ×0.2 / ×1 | $0.4 / $2 | $0.04 / $0.2 | |
| Codestral | ×0.15 / ×0.45 | $0.3 / $0.9 | $0.03 / $0.09 |
FAQ
How is the price counted?
You buy a pack of balance tokens. Each model charges its own rate for input and output tokens, shown in the “Balance rate” column. For Claude Sonnet 5.5 it is ×1 for input and ×5 for output.
Can I switch between models?
Yes. The balance is shared, and the model is picked per request with the model field or in the cabinet chat.
Why are prices below the official ones?
KLINCH has no subscription and one discount for every model: the price is about a tenth of the official one across the board.
Which models are live right now?
GET /v1/models with your key returns the current list, and the status page shows how the models work in real time.
One key, one balance, every top model
Claude, GPT, Gemini, DeepSeek and Grok at a tenth of the official price. No subscription, pay as you go.