MMGA Token Usage
How staked MMGA influences post-launch routing, node reliability, the community model marketplace and governance.
MMGA is the community’s core network token. Its utility connects compute supply, reliable inference, community video models and governance into one economic system.
The community’s core token
MakeMiningGreatAgain (MMGA) is designed as the coordination asset of the MineAPI community—not as a replacement for real inference revenue.
Inference Credit pays device owners for completed user workloads. MMGA adds the long-term coordination layer: it aligns operators, rewards reliable capacity, supports the community model economy and gives stakeholders a voice in network decisions.
The four primary MMGA utilities are:
- Stake-weighted job routing after the Launch Phase.
- Reliability rewards and penalties for node behavior.
- Burn-to-List access and the transaction medium for community fine-tuned video models.
- Community proposals and model selection.
1. Stake-weighted job routing
During the first 180-day Launch Phase, MineAPI may prioritize security, benchmarks and direct network controls while the operator system matures.
After the Launch Phase, the amount of MMGA staked by an eligible account becomes one input to the Router’s job assignment score. When two nodes satisfy the technical and reliability requirements of a job, an account with more actively staked MMGA can receive a higher probability of being assigned higher-value inference work.
Eligible routing weight
= verified device performance
× reliability score
× effective availability
× post-launch MMGA stake factor
Staking does not allow an incompatible or unreliable node to buy its way into a job. Model support, memory, runtime, latency, region and historical quality remain mandatory routing gates.
The intended effect is to reward operators who commit both useful compute and long-term economic stake to the network.
| Routing input | Role |
|---|---|
| Technical compatibility | Determines whether a node can run the job |
| Reliability and quality | Protects users from failed or fraudulent inference |
| Effective availability | Measures capacity actually offered to the network |
| Staked MMGA | Adds post-Launch-Phase economic routing weight |
2. Reliability rewards and penalties
Highly available nodes receive a higher proportion of the eligible MMGA network reward budget. Availability alone is not enough: the node must also complete assigned work successfully and behave honestly.
The reliability system considers signals such as:
- effective online time;
- accepted-job completion rate;
- inference validation success;
- timeout and disconnect frequency;
- fraudulent or manipulated output evidence;
- repeated mismatch between advertised and measured capability.
Rewarding high availability
Nodes that remain eligible, responsive and successful can receive a higher reward weight within an emission epoch.
Epoch reward weight
= verified inference performance
× effective online time
× reliability multiplier
× network utilization multiplier
Penalizing fraud and repeated failure
If a node is found to commit fraud, manipulate results, misreport capability or repeatedly fail inference jobs, MineAPI can reduce the MMGA that would otherwise be issued to that node.
The penalty is deducted from unissued and pending MMGA rewards. It is not represented as an automatic confiscation of tokens already settled in a user’s wallet.
Depending on severity, enforcement can include:
- reduced reward multiplier;
- forfeiture of pending, not-yet-issued MMGA;
- temporary routing suspension;
- removal from higher-value job eligibility;
- node or account exclusion for verified fraud.
This design keeps capacity online while making persistent failure and dishonest behavior economically unattractive.
3. Community model marketplace
MMGA is the primary listing and transaction token for the MineAPI community fine-tuned video-model marketplace.
Community developers can fine-tune models such as MiniMax H3 for a specific visual style, motion pattern, commercial workflow or domain. MineAPI provides the inference capacity and API delivery layer; the model creator supplies the fine-tuned model and its usage terms.
Burn to list
To list a model project on the marketplace, the seller burns a defined amount of MMGA.
Seller submits model and evaluation package
→ Technical and policy review
→ Seller burns required MMGA listing amount
→ Model becomes eligible for marketplace listing
→ MineAPI provides distributed inference capacity
Burning MMGA creates a real cost for listing, reduces spam and signals that the seller is committed to maintaining the model. Burned MMGA is permanently removed from transferable supply.
Passing the burn requirement does not override technical review, safety requirements, intellectual-property checks or community policy.
MMGA as the model transaction medium
After a community model is approved and listed, sellers can price the model, license or marketplace access right in MMGA. Buyers use MMGA to purchase that right, making MMGA the settlement medium between community model creators and users.
MMGA marketplace transactions can represent:
- acquisition of a fine-tuned model package;
- a commercial or limited-use model license;
- access to a private model endpoint;
- a seller-defined usage entitlement;
- community revenue-sharing attached to a model project.
Buyer pays model price in MMGA
→ Marketplace verifies the purchase or license
→ Seller receives the defined MMGA proceeds
→ Buyer receives the model access entitlement
The model transaction and the inference workload are separate economic items. A marketplace purchase can be settled in MMGA while the compute consumed when running the model is billed independently as MineAPI Credit, unless a future listing explicitly bundles inference capacity.
Marketplace rules may allocate a published portion of each transaction to the seller, the community treasury, protocol fees or an additional burn mechanism. Exact fee splits must be visible before the buyer confirms the transaction.
Marketplace roles
| Participant | Contribution |
|---|---|
| Model seller | Fine-tuned model, documentation and maintenance |
| MineAPI | Evaluation workflow, routing, inference capacity and API delivery |
| Device hosts | Compatible compute for model inference |
| Model buyer | MMGA payment for the model, license or access entitlement |
| Community | Stake-weighted discovery and model selection |
Final marketplace revenue-sharing and burn amounts will be published with the production marketplace schedule.
4. Governance and model selection
Staked MMGA provides voting power for community coordination.
Community proposals
Eligible participants can use staked MMGA to signal support for MineAPI Improvement Proposals covering areas such as:
- minimum MMGA commitments for network operators;
- benchmark and verification rules;
- reward and utilization parameters;
- community treasury priorities;
- new runtime or device support;
- regional capacity programs.
Voting power is derived from eligible staked MMGA under the applicable proposal rules. Quorum, voting periods and execution requirements can vary by proposal type.
Model selection
Community members can assign staked-MMGA voting power to MiniMax H3 fine-tune projects in the Model Leaderboard.
Votes help determine which models should advance into evaluation, receive community support or become candidates for MineAPI network deployment. Voting is a curation signal, not an automatic technical approval.
Utility flow
Stake MMGA
├─ Increase post-launch routing weight for eligible reliable nodes
├─ Receive community proposal voting power
└─ Vote on community video-model selection
Operate reliable nodes
├─ Earn inference Credit from real completed jobs
└─ Earn a higher share of eligible MMGA rewards
Publish a community model
├─ Pass evaluation and policy review
├─ Burn MMGA to list
├─ Sell the model, license or access entitlement in MMGA
└─ Use MineAPI community compute for inference delivery
Current implementation status
The website currently demonstrates staking balances, 14-day locking, model voting and community proposals in the browser.
The following mechanisms are not yet enforced by a production backend or smart contract:
- stake-weighted Router assignment;
- automatic reliability scoring and reward penalties;
- MMGA burn transactions;
- MMGA model purchases and seller settlement;
- model marketplace settlement;
- binding on-chain governance execution.
Until those systems launch, displayed votes, balances, token prices, pool actions and proposal results remain frontend Mock data. The production contract address and exact protocol parameters must be published through official MineAPI channels.