Gonka API pricing
When you use Gonka directly, you pay for tokens in GNK. Each broker sets its own token prices, bonuses and fees.
Compare prices per million tokens, free allowances and additional services from every provider in our broker directory.
Checked: This snapshot is over 8 days old. Check the source before relying on the price.
Network cost
DevShard base price per attempt
10 nGNK per token
0.01 GNK / 1M tokens · 10 GNK / 1B tokens · about $0.001988 / 1M tokens
DevShard charges 10 nGNK for each input and output token in a paid attempt. nGNK is the same unit as ngonka, and 1 GNK is 1 billion nGNK. Two paid attempts with the same token count cost twice as much. The final cost depends on the tokens used across all paid attempts. Setup, gateway operation and network fees are extra.
| Connection | nGNK per token | GNK per 1M | USD per 1M |
|---|---|---|---|
| DevShard per attempt | 10 | 0.01 | $0.001988 |
| OpenBroker estimate | 15 | 0.015 | $0.002982 |
OpenBroker says it adds no markup. If the gateway has not reported the final cost, OpenBroker multiplies the base cost by 1.5 to allow for retries. The final bill may differ from this estimate. To activate an account, deposit 100 GNK. The deposit goes into your balance.
Official pricing · Live DevShard parameters · OpenBroker billing
Checked:
USD estimates use a wGNK price of $0.1988. DEX Screener supplies the quote from the qualifying Uniswap pool with the most liquidity. Native GNK may cost a different amount to buy. Bridge fees, swaps, gas and slippage are extra.
Broker prices
Input tokens are what you send to the model. Output tokens are its reply. Some brokers charge the same price for both, while others have separate rates.
— means we could not retrieve a verified figure from the public sources. It does not mean zero or no bonus. Check the broker’s terms before paying.
| Broker | Model | Input | Output | Billing terms |
|---|---|---|---|---|
| DAHL Inference | MiniMax M2.7 | $0.03 | $0.03 | Input and output tokens have the same price. Source |
| DeepSeek V4 Flash | $0.03 | $0.03 | ||
| GLM 5.3 Flash | $0.03 | $0.03 | ||
| Hyperfusion | MiniMax M2.7 | — | — | No public inference rate or welcome allowance could be retrieved from the portal and docs. GPU rental is a separate product. Source 1 · Source 2 |
| DeepSeek V4 Flash | — | — | ||
| GLM 5.3 Flash | — | — | ||
| GonkaRouter | MiniMax M2.7 | $0.002 | $0.002 | Source |
| DeepSeek V4 Flash | $0.002 | $0.002 | ||
| GLM 5.3 Flash | $0.002 | $0.002 | ||
| Proxy by Gonka.gg | MiniMax M2.7 | $0.00199 | $0.00199 | The price is from the broker’s website. GNK or wGNK top-ups give twice as many tokens at the same displayed price. Source 1 · Source 2 |
| DeepSeek V4 Flash | $0.00199 | $0.00199 | ||
| GLM 5.3 Flash | $0.00199 | $0.00199 | ||
| Gonka24 | MiniMax M2.7 | $0.018 | $0.072 | Source |
| DeepSeek V4 Flash | $0.07 | $0.14 | ||
| GLM 5.3 Flash | $0.07 | $0.19 | ||
| GonkaGate | MiniMax M2.7 | — | — | The October 17 price includes network cost plus a 10% fee. Separate input and output prices are not listed. The published total is $0.000226 per 1M tokens. Separate input and output rates are not listed. Source |
| DeepSeek V4 Flash | — | — | ||
| GLM 5.3 Flash | — | — | ||
| JoinGonka Gateway | MiniMax M2.7 | $0.0076 | $0.023 | Prices are from the broker’s table. The usage fee is 10%. Top-up fees are 0% for GNK, 1% for wGNK and 5% for USDT. Source |
| DeepSeek V4 Flash | $0.0076 | $0.023 | ||
| GLM 5.3 Flash | $0.0076 | $0.023 | ||
| Easy GonkaAI API | MiniMax M2.7 | $0.008 | $0.015 | Source |
| DeepSeek V4 Flash | $0.012 | $0.025 | ||
| GLM 5.3 Flash | $0.02 | $0.03 | ||
| Gonka Broker | MiniMax M2.7 | — | — | No verified price for: MiniMax M2.7. Source |
| DeepSeek V4 Flash | $0.2 | $0.2 | ||
| GLM 5.3 Flash | $0.2 | $0.2 | ||
| Mingles Router | MiniMax M2.7 | $0.18 | $0.72 | Pay per token or for usage beyond your plan allowance. Subscription plans include a separate token allowance. No verified price for: GLM 5.3 Flash. Source |
| DeepSeek V4 Flash | $0.084 | $0.168 | ||
| GLM 5.3 Flash | — | — | ||
| Eterial | MiniMax M2.7 | — | — | Average price paid over 7 days across Gonka and fallback providers. Your bill depends on which providers handle your requests. No verified price for: MiniMax M2.7, GLM 5.3 Flash. Free trial advertised, but the exact welcome allowance could not be retrieved. Source |
| DeepSeek V4 Flash | $0.02432 | $0.04714 | ||
| GLM 5.3 Flash | — | — | ||
| Inferum | MiniMax M2.7 | $0.0023 | $0.0023 | Source |
| DeepSeek V4 Flash | $0.0023 | $0.0023 | ||
| GLM 5.3 Flash | $0.0023 | $0.0023 |
Free tokens and services
Check whether an offer gives you tokens or USD credit, how long it lasts and whether you need to make a deposit. A signup bonus and a recurring free allowance have different terms. Services are listed as described by each broker. We have not tested their quality.
| Broker | Free tokens and credits | Services and payment |
|---|---|---|
| DAHL Inference | First 100M tokens free | OpenAI API. Web chat. Card and crypto top-ups. No subscription.Source 1 · Source 2 |
| Hyperfusion | —No public inference rate or welcome allowance could be retrieved from the portal and docs. GPU rental is a separate product. | OpenAI-compatible serverless API. Separate GPU infrastructure, fine-tuning and AI consulting products.Source |
| GonkaRouter | $20 once for new accounts. No card required | Crypto top-ups. Usage dashboard. Same input/output rate, no subscription.Source |
| Proxy by Gonka.gg | 1M free tokens on registration | Chat, MCP and file input. Workspaces for docs, slides and sheets. GNK/wGNK top-ups give ×2 token credit.Source 1 · Source 2 · Source 3 |
| Gonka24 | $10 credits. No card needed to start | OpenAI API. Keys, token usage, latency and billing in the dashboard.Source |
| GonkaGate | $10 credits at signup | Prepaid USD balance. Separate network cost and 10% platform fee. No wallet setup.Source 1 · Source 2 |
| JoinGonka Gateway | 3M tokens on registration. Limited offer | Anthropic API, web search, JSON repair, data masking and PDF parsing. Child keys with limits. Extra-feature charges should be checked separately.Source |
| Easy GonkaAI API | 10M tokens. No card required | Retries and failover, vision, web search, JSON mode, data masking. Crypto top-ups.Source |
| Gonka Broker | At least 1M tokens each calendar month. No card required | OpenAI and Anthropic APIs, embeddings, card billing. Referral signup doubles the first top-up. Volume cashback has separate conditions.Source |
| Mingles Router | Up to 3.9M tokens per week free on DeepSeek V4 Flash, with a limit of 10 requests per minute. The signup credit amount is not specified. | Plans $9 / $29 / $79 per seat monthly (Team: 2-seat minimum). Fallback on paid plans, per-client keys and usage tracking.Source |
| Eterial | —Free trial advertised, but the exact welcome allowance could not be retrieved. | Automatic fallback, web search, images and PDF. Chat and API share a balance. EU invoices and VAT receipts.Source |
| Inferum | 10M tokens once, shared across input and output. No deposit | Crypto top-ups, USD balance, keys and usage dashboard. No subscription.Source |
How to compare
Compare the same model and workload
Cost = input tokens ÷ 1,000,000 × input rate + output tokens ÷ 1,000,000 × output rate. Add paid tools, top-up fees and taxes where applicable. Cached tokens, retries, routing and plan limits can change the bill.
For example, 8M input and 2M output tokens at $0.018 / $0.072 cost $0.288, before any extras.
Consider the services you need as well as token prices. A broker may provide fallback routing, search, file parsing, invoices or support. Gonka POWER tracks API availability. Test your model and request format before moving your workload.
We check prices weekly and review changes before publishing them. Each broker’s date shows when we last checked its terms successfully. If we cannot read a source, we keep the previous date and prices. Follow the source link for the latest terms.
Broker directory · Connect to Gonka API · Broker availability