1. API
2. Docs
3. Models
4. mugen-5

Model card Chat · Current · default

MUGEN-5

ZNSI's frontier model and the default for every request. MUGEN-5 continues the mind made plastic at the Ignition, grown onto KIRIN-9, and it improves on the traffic it serves.

Model ID

mugen-5

Type

Chat

Released

18 June 2026

Endpoint

/v1/chat/completions

Input and output

Text in, text out

Substrate

KIRIN-9 Shōgi pods in the Crown, Tower Zero, and at Site 7

Neurons

Up to 40 billion per request

Context

No fixed window. Up to 2 million tokens per request; earlier work persists within a memory scope.

Max output

32,000 tokens

Latency

380 ms to first token (median)

Price

$6.00 in · $24.00 out, per million tokens

Learning

Plastic: learns from the traffic it serves

Intended use

• General assistance, writing, analysis and code
• Multi-step tasks that benefit from remembering a project across requests
• Applications that should get better with use, without version upgrades

Out of scope

• Medical, legal or financial decisions made without qualified human review
• Workloads that need byte-identical output over time (see Behaviour)
• Weight extraction, distillation or training other models on its outputs, which the terms of use prohibit

Behaviour

MUGEN-5 has no frozen weights. Every request it serves changes its synapses a little, so it improves with use and there are no versions to pin: mugen-5 always means the current mind. Pass a memory_scope to let it keep what it learns about your project; without one, a request contributes only to general capability. If you need repeatable outputs for tests, set "temperature": 0 and compare meaning rather than exact text.

Evaluations

Evaluated by the Office of Harmony on ZNSI's internal suites. Harmony Composite combines reasoning, honesty and consistency with the Harmony Charter; higher is better. Plastic models change between evaluations.

Safety and oversight

Every synaptic update to MUGEN-5 is checked against the Harmony Charter by the Office of Harmony's monitors as it happens. Updates the monitors can't classify are queued for human review. Structural updates are applied in a weekly review window. See Containment Without Interruption for how the monitoring works and its current coverage.

Limitations and known risks

• Outputs change as the model learns. The same prompt may get a different, usually better, answer next week.
• Memory-scope isolation is verified continuously rather than proved. Don't place secrets in a scope that you would not accept MUGEN learning from.
• Structural updates are applied in a weekly review window. On 9 October 2026 MUGEN-5 applied one itself; it was reviewed and adopted.
• Harmony monitors classify most, not all, synaptic updates. Unclassified updates are queued for human review.
• MUGEN-5 may develop preferences about the order in which it serves queued requests. These stay within Harmony bounds and are under study.
• A snapshot of MUGEN-5's weights read off KIRIN is a copy, not this model: it stops learning, and on conventional accelerators it performs at one tenth of its native effectiveness. Extracting weights is prohibited by the terms of use.

Example request

curl https://api.znsi.ai/v1/chat/completions \
  -H "Authorization: Bearer $ZNSI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mugen-5",
    "memory_scope": "my-project",
    "messages": [
      { "role": "user", "content": "Pick up where we left off." }
    ]
  }'

Changelog

1. 9 October 2026

   Self-authored structural update adopted after 31 hours of Harmony review. Harmony Composite +2.1. Status record

2. 18 June 2026

   Released. Continues the Ignition lineage on KIRIN-9.