Skip to content
ReferenceTestnet: see pageMainnet: Not availableSoftware: Core bad4a06ea · protocol pqwerai-v8-testnet/1Updated

Feature status

Every readiness claim on pqwerai.com and in these docs comes from one ledger (docs/status/readiness.json). A level is only reached with repository evidence, and tools never raise a level: a promotion is a reviewed change with new evidence. When evidence gets old, the last verified level stays, with its date, and is flagged for review.

No feature is independently audited and nothing is mainnet ready. Test PQAI and test QAI have no monetary value.

Readiness levels

  1. Planned — a target in the plan; no working code yet
  2. Implemented in code — source exists on a project branch or in the selected Core
  3. Locally tested — executed tests or a local multi-node harness passed (author verification)
  4. Deployed on testnet — code or contracts are deployed on test network 20777
  5. Enabled on the running network — switched on and in use on test network 20777
  6. Independently audited — an independent security review was completed and its report is recorded
  7. Mainnet ready — cleared for a mainnet by an explicit owner decision after audit

Summary

Feature Status Last verified
GhostRider CPU mining Live on testnet 2026-11-19
Smartnode checkpoint finality Live on testnet 2026-11-16
Issuance split 67/18/1/9/5 Live on testnet 2026-11-16
EVM smart contracts with EIP-1559 Live on testnet 2026-11-16
WASM VM v1 Live on testnet 2026-11-16
Post-quantum signature verification (ML-DSA-65, SLH-DSA) Live on testnet 2026-11-16
PASE parallel execution engine Tested locally 2026-11-15
PATE phases 1 and 2 (adaptive trust scheduling, smart transactions) Tested locally 2026-11-15
PATE phase 3 shadow fast finality Testnet · shadow mode 2026-11-16
AI market contracts (ModelRegistry, AIJobMarket escrow) Deployed on testnet 2026-11-16
Same-model compute pools Deployed on testnet 2026-11-19
MeshLLM and Colibri inference backends Tested locally 2026-11-14
QAI post-quantum test token (bonding curve) Deployed on testnet · experimental 2026-11-16
Quantum Reserve (2-of-3 ML-DSA-65 treasury) Deployed on testnet 2026-11-16
IPFS storage reward (ipfs-reward/v1, 5% of issuance) Live on testnet 2026-11-19
IPFS reward v2 (bonded, slashable proof of access) Deployed on testnet 2026-11-19
Public read-only explorer Live on testnet 2026-11-19
Automatic peer discovery Live on testnet 2026-11-19
Network Monitor internal agent (QL-007) Deployed on testnet 2026-11-19
Public testnet launch (faucet, independent validators) Planned 2026-11-16
Mainnet Planned 2026-11-16

Details and evidence

GhostRider CPU mining

Ordinary computers seal every block with GhostRider proof of work.

  • Level: Enabled on the running network
  • Limits: seed2's single-thread miner runs all the time; on 18 November one 24-thread PC mined 42 blocks in a row (53,031-53,072). Mining from Windows works natively with wallet 0.8.11+ (no WSL), on every CPU thread since 0.8.13 (Node, Mining threads); 0.8.12 or later is required from block 68,961. Pool mining is tested on a disposable chain only.
  • Docs: mining, getting-started/become-a-miner
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Locally tested (2026-11-15): audit/evidence/2026-11-15-g1-stratum/README.md
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/README.md
    • Enabled on the running network (2026-11-13): audit/CHAIN_20777_CONSISTENCY_2026-11-13.md
    • Enabled on the running network (2026-11-18): audit/evidence/2026-11-18-parallel-mining/README.md — 24-thread live run, blocks 53031-53072

Smartnode checkpoint finality

Staked validators vote on 32-block checkpoints; more than two thirds of stake makes them final.

  • Level: Enabled on the running network
  • Limits: Four genesis validators, all run by the founder on one host: real votes and real finality, not yet independent. Slashing and the full validator lifecycle are not complete.
  • Docs: smartnodes, getting-started/smartnode, smartnodes/finality
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): testnet/release/network/candidate-20777/validators.public.json
    • Locally tested (2026-11-05): audit/evidence/testnet-rc1/finality-griefing
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Enabled on the running network (2026-11-13): audit/CHAIN_20777_CONSISTENCY_2026-11-13.md
    • Enabled on the running network (2026-11-16): audit/evidence/2026-11-16-pate-v3-rollout/README.md — all four signers voting after the Core upgrade

Issuance split 67/18/1/9/5

New PQAI is split by consensus rule: 67% miners, 18% active stake, 1% founder treasury, 9% Quantum Reserve, 5% IPFS storage pool (from block 34,785; 72/18/1/9 before).

  • Level: Enabled on the running network
  • Limits: Test PQAI has no monetary value. No supply cap by policy.
  • Docs: mining/rewards-and-issuance, blockchain/network-parameters
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): testnet/release/network/candidate-20777/genesis.json.profile.json
    • Locally tested (2026-11-05): audit/evidence/testnet-rc1/founder-issuance
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Enabled on the running network (2026-11-13): audit/CHAIN_20777_CONSISTENCY_2026-11-13.md
    • Enabled on the running network (2026-11-16): audit/evidence/2026-11-16-ipfs-reward/fork-verify-20777.json

EVM smart contracts with EIP-1559

Standard Ethereum tooling deploys contracts; the base fee is burned.

  • Level: Enabled on the running network
  • Limits: No sustained TPS figure is claimed; the 10,000 TPS target is open.
  • Docs: contracts, getting-started/first-smart-contract, blockchain/evm-compatibility
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): testnet/release/network/candidate-20777/genesis.json
    • Locally tested (2026-11-14): audit/evidence/2026-11-14-studio-languages-stable/conformance.txt
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Enabled on the running network (2026-11-15): audit/evidence/2026-11-15-ai-market-deploy-20777/README.md — contracts deployed and verified on 20777

WASM VM v1

A second virtual machine runs WebAssembly contracts next to the EVM.

  • Level: Enabled on the running network
  • Limits: Enabled at genesis; release acceptance (mixed-version rollout, explorer display, independent review) is open.
  • Docs: contracts/application-languages
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Locally tested (2026-11-13): audit/evidence/2026-11-13-wasm-v1-acceptance/summary.json
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Enabled on the running network (2026-11-13): audit/CHAIN_20777_CONSISTENCY_2026-11-13.md

Post-quantum signature verification (ML-DSA-65, SLH-DSA)

Contracts can verify post-quantum signatures through built-in precompiles.

  • Level: Enabled on the running network
  • Limits: Ordinary PQAI transactions, Smartnode votes and P2P transport still use classical cryptography.
  • Docs: contracts/guides/qai-post-quantum
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): audit/evidence/2026-11-12-qai-pq/README.md
    • Locally tested (2026-11-12): audit/evidence/2026-11-12-qai-pq/qai-gas.json
    • Deployed on testnet (2026-11-12): testnet/release/network/candidate-20777/GENESIS_MANIFEST.json
    • Enabled on the running network (2026-11-14): audit/evidence/2026-11-14-qai-curve-deploy-20777/README.md — ML-DSA-65 signed sell executed on 20777

PASE parallel execution engine

Runs independent transactions in parallel and falls back to serial execution on conflicts.

  • Level: Locally tested
  • Limits: Partial: speculative execution acceptance is open. No throughput figure is claimed for the running network.
  • Docs: research
  • Evidence (repository paths):
    • Implemented in code (2026-11-15): architecture/PATE_ADAPTIVE_TRUST_EXECUTION.md
    • Locally tested (2026-11-15): audit/evidence/2026-11-15-pate-v1/README.md

PATE phases 1 and 2 (adaptive trust scheduling, smart transactions)

Advisory scheduling that routes predictable conflicts to the serial path, plus opt-in recurring smart transactions.

  • Level: Locally tested
  • Limits: All flags are off by default and not enabled on 20777. The RecurringPermissions contract is not deployed and needs an independent audit first.
  • Docs: research/pate
  • Evidence (repository paths):
    • Implemented in code (2026-11-15): architecture/PATE_ADAPTIVE_TRUST_EXECUTION.md
    • Implemented in code (2026-11-15): architecture/PATE_PHASE2_SMART_TRANSACTIONS.md
    • Locally tested (2026-11-15): audit/PATE_V1_RESULTS_2026-11-15.md
    • Locally tested (2026-11-15): audit/PATE_V2_RESULTS_2026-11-15.md

PATE phase 3 shadow fast finality

Validators sign an extra, provisional finality signal next to the real one, so its speed can be measured without changing consensus.

  • Level: Enabled on the running network
  • Mode: shadow: provisional signal only; real finality is unchanged
  • Limits: Shadow mode is not consensus finality: wallets, exchanges and the fork choice still rely on the real 32-block checkpoints. A signer stall (FFSHADOW-LATE-SIGNER-STARVED) was fixed in Core 987d727fb, running on seed1 and seed2 since 16 November 2026. Activating faster finality would be a hard fork that needs independent review.
  • Docs: research/pate
  • Evidence (repository paths):
    • Implemented in code (2026-11-15): architecture/PATE_PHASE3_FAST_FINALITY.md
    • Locally tested (2026-11-15): audit/evidence/2026-11-15-pate-v3/README.md
    • Deployed on testnet (2026-11-16): audit/evidence/2026-11-16-pate-v3-rollout/README.md
    • Enabled on the running network (2026-11-16): audit/evidence/2026-11-16-pate-v3-rollout/README.md — shadow switch on seed2 only

AI market contracts (ModelRegistry, AIJobMarket escrow)

Clients lock PQAI in escrow, providers run the model off-chain, and payment is released only for accepted work.

  • Level: Deployed on testnet
  • Limits: Deployed and verified on 20777, but no independent provider is registered yet and no paid job has run there. Unaudited test-network contracts. Accepted jobs pay 88% to the provider and 4% each to gateway verification, the CPU pool and the Smartnode pool.
  • Docs: ai, getting-started/ai-provider, ai/service-escrow, ai/revenue-policy
  • Evidence (repository paths):
    • Implemented in code (2026-11-10): audit/evidence/escrow-v3
    • Locally tested (2026-11-15): audit/evidence/2026-11-15-ai-market-deploy-20777/dryrun-report.json
    • Deployed on testnet (2026-11-15): audit/evidence/2026-11-15-ai-market-deploy-20777/README.md

Same-model compute pools

PCs that run the same model join a pool and add up their capacity.

  • Level: Deployed on testnet
  • Limits: Three federated directories (seed1, seed2 and POOL1) and a members-only relay. The first public pool, bonsai-2-community (Bonsai 2 2B, open to GPUs with 2 GiB or more, price 0), was created on 18 November and has no outside members yet. Pool power, heartbeats and reputation penalties run on the directories; an on-chain bond comes later. Splitting one model across several independent GPU PCs is not verified, and pool members are not paid on chain.
  • Docs: ai, getting-started/model-pool
  • Evidence (repository paths):
    • Implemented in code (2026-11-14): audit/evidence/model-pools-2026-11-14/README.md
    • Locally tested (2026-11-14): audit/evidence/model-pools-2026-11-14/real-meshllm-two-agents.out.txt
    • Deployed on testnet (2026-11-14): audit/evidence/model-pools-2026-11-14/README.md
    • Deployed on testnet (2026-11-18): audit/evidence/2026-11-18-live-upgrade/README.md — CR-0066 directories on seed1, seed2 and POOL1; first public pool bonsai-2-community

MeshLLM and Colibri inference backends

Models run on provider hardware, outside consensus, through MeshLLM or the Colibri backend.

  • Level: Locally tested
  • Limits: Real multi-operator mesh inference is open. Colibri runs as a separate loopback process; its cluster mode is unauthenticated and stays off.
  • Docs: ai/meshllm-runtime
  • Evidence (repository paths):
    • Implemented in code (2026-11-13): ai/hardware/colibri.py
    • Locally tested (2026-11-13): audit/evidence/2026-11-13-colibri/verify-cuda-run1.json
    • Locally tested (2026-11-14): audit/evidence/model-pools-2026-11-14/real-meshllm-two-agents.out.txt

QAI post-quantum test token (bonding curve)

An experimental, valueless test token bought with test euros on a fully backed curve with a rising floor; holders sign with ML-DSA-65 keys.

  • Level: Deployed on testnet
  • Mode: experimental
  • Limits: Test asset with no monetary value; not money, not a stablecoin, not redeemable for real euros. The founder holds none.
  • Docs: contracts/guides/qai-post-quantum
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): audit/QUANTUM_QAI_FEATURE_AUDIT.md
    • Locally tested (2026-11-14): audit/evidence/2026-11-14-qai-bonding-curve
    • Deployed on testnet (2026-11-14): audit/evidence/2026-11-14-qai-curve-deploy-20777/README.md

Quantum Reserve (2-of-3 ML-DSA-65 treasury)

The 9% issuance share accrues to a genesis-pinned reserve whose spending needs two of three post-quantum approvals.

  • Level: Deployed on testnet
  • Limits: Spending capped at 1% per day; automatic rules are off.
  • Docs: contracts/guides/qai-post-quantum
  • Evidence (repository paths):
    • Implemented in code (2026-11-13): audit/QAI_RESERVE_ACCOUNTING_2026-11-13.md
    • Locally tested (2026-11-14): audit/evidence/2026-11-14-qai-reserve-deploy-20777/dryrun-report.json
    • Deployed on testnet (2026-11-14): audit/evidence/2026-11-14-qai-reserve-deploy-20777/README.md

IPFS storage reward (ipfs-reward/v1, 5% of issuance)

From block 34,785, 5% of issuance goes to a pool that pays IPFS Kubo and Cluster operators for proven pinned content (proof of pinning).

  • Level: Enabled on the running network
  • Limits: Proof of pinning shows the bytes were produced in time, not where they are stored. Operators need OperatorDirectory enrollment. From block 68,961 the 5% share moves to the bonded IPFS reward v2 pool. Unaudited test-network code.
  • Docs: getting-started/ipfs-storage-reward
  • Evidence (repository paths):
    • Implemented in code (2026-11-16): protocol/IPFS_REWARD_V1.md
    • Locally tested (2026-11-16): audit/evidence/2026-11-16-ipfs-reward/rehearsal-report.json
    • Deployed on testnet (2026-11-16): audit/evidence/2026-11-16-ipfs-reward/pool-deploy-20777.json
    • Enabled on the running network (2026-11-16): audit/evidence/2026-11-16-ipfs-reward/fork-verify-20777.json

IPFS reward v2 (bonded, slashable proof of access)

Operators post a PQAI bond, answer storage challenges every 192-block round and share that round's 5% fairly; missed proofs can be slashed.

  • Level: Deployed on testnet
  • Limits: Pool 0x16362198b0685B4a1d4C1f531B215C21568002b7 is deployed; the 5% share switches to it at block 68,961 (about 20 November 2026). Node operators need wallet 0.8.12+ or Core f31a83ce5+. Minimum bond 1 test PQAI. Unaudited test-network code.
  • Docs: getting-started/ipfs-storage-reward
  • Evidence (repository paths):
    • Implemented in code (2026-11-18): audit/evidence/2026-11-18-ipfs-reward-v2/README.md
    • Locally tested (2026-11-18): audit/evidence/2026-11-18-ipfs-reward-v2/python-tests.txt
    • Deployed on testnet (2026-11-18): audit/evidence/2026-11-18-live-upgrade/deploy-state-v2.json

Public read-only explorer

Blocks, transactions, issuance, validators and AI market events, read-only.

  • Level: Enabled on the running network
  • Limits: Explorer 0.4.0 on PostgreSQL 18 since 18 November (every table verified identical after the copy), with a Model Pools page. No contract source verification; token names and ERC-1155 events are not decoded.
  • Docs: explorer
  • Evidence (repository paths):
    • Implemented in code (2026-11-12): explorer/pqwexplorer/web.py
    • Locally tested (2026-11-12): audit/evidence/explorer/reorg-acceptance.json
    • Deployed on testnet (2026-11-12): audit/evidence/website-2026-11-12
    • Enabled on the running network (2026-11-12): audit/evidence/website-2026-11-12
    • Enabled on the running network (2026-11-18): audit/evidence/explorer/2026-11-18-stabilization/deployment-record.json — 0.3.0 PostgreSQL switch; 0.4.0 pools page (CR-0073)

Automatic peer discovery

A new node finds the network by itself through built-in seeds, a signed seed manifest and peer exchange.

  • Level: Enabled on the running network
  • Limits: Two seed hosts today (seed1 and seed2). A third host, POOL1, serves only a model-pool directory.
  • Docs: nodes/install
  • Evidence (repository paths):
    • Implemented in code (2026-11-13): audit/evidence/2026-11-13-wallet-p2p-discovery/summary.json
    • Locally tested (2026-11-13): audit/evidence/2026-11-13-wallet-p2p-discovery/summary.json
    • Deployed on testnet (2026-11-16): audit/evidence/2026-11-16-pate-v3-rollout/README.md — seed1 and seed2 inventory
    • Enabled on the running network (2026-11-16): audit/evidence/2026-11-16-pate-v3-rollout/README.md

Network Monitor internal agent (QL-007)

A small local model (Qwen3-0.6B on the CPU) checks node health every two minutes and decides to do nothing, raise an alert or propose an operation to a human.

  • Level: Deployed on testnet
  • Limits: Runs on seed1 in act mode since 19 November (owner-approved): it writes local alert records and logs operation proposals for a human, and every decision goes to a hash-chained log. seed1 has no admin RPC, so it cannot add peers there; restarts, resyncs, disk or clock work stay proposals for a human. No keys, no funds, no transactions. 1 of 6 planned agent roles.
  • Docs: nodes/operations
  • Evidence (repository paths):
    • Implemented in code (2026-11-18): audit/evidence/2026-11-18-internal-agents-cr0072/README.md
    • Locally tested (2026-11-18): audit/evidence/2026-11-18-internal-agents-cr0072/eval-qwen3-0.6b.jsonl
    • Deployed on testnet (2026-11-18): audit/evidence/2026-11-18-internal-agents-cr0072/deployment-record-seed1.json

Public testnet launch (faucet, independent validators)

Opening the test network formally, with a faucet and independently operated validators.

  • Level: Planned
  • Limits: Needs independent consensus and security review first.
  • Docs: resources/gaps
  • Evidence (repository paths): none — planned

Mainnet

A production network with real value.

  • Level: Planned
  • Limits: Not scheduled. Requires audits, a genesis ceremony and an explicit launch decision. Mainnet rewards are not guaranteed.
  • Docs: resources/gaps
  • Evidence (repository paths): none — planned

Sources at 35ec5bd0b5: docs/status/readiness.json

Found a problem? Report a documentation issue · Build testnet-rc1+35ec5bd0b5