Profile Model Fleet — Three-Provider Assignments
The fleet uses three providers:
- Xiaomi MiMo (
provider: xiaomi, base_url: https://token-plan-sgp.xiaomimimo.com/v1) — primary heavy lifter, 1B token giveaway
- DeepSeek Direct (
provider: deepseek, base_url: https://api.deepseek.com) — premium reasoning for research/market analysis
- Nous Portal (
provider: nous, base_url: https://inference-api.nousresearch.com/v1) — free tier for lightweight workers (deepseek-v4-flash:free)
- OpenRouter (
provider: openrouter) — used by secretary for owl-alpha
Profile → Model Map (as of 2026-05-27)
Senna (default chat session)
|
|
| Model |
mimo-v2.5-pro |
| Provider |
xiaomi |
| Base URL |
https://token-plan-sgp.xiaomimimo.com/v1 |
Oracle (market intelligence)
|
|
| Model |
deepseek-v4-pro |
| Provider |
deepseek |
| Base URL |
https://api.deepseek.com |
NOTE (2026-05-27): Researcher was briefly switched to xiaomi/mimo-v2.5-pro due to missing DEEPSEEK_API_KEY in its profile .env. Key was copied from root .env, researcher switched back to deepseek-v4-pro with reasoning_effort: high.
Tier 1 — Lightweight (Xiaomi + OpenRouter, reasoning: none)
Fast, cheap workers for routing, triage, light tasks.
| Profile |
Provider |
Model |
Role |
| Foreman |
xiaomi |
mimo-v2.5-pro |
Swarm orchestrator — decomposes tasks, routes to specialists |
| Secretary |
openrouter |
owl-alpha |
Inbox triage, docs, writing |
| DevOps |
nous |
deepseek-v4-flash:free |
Ops health checks, dependency updates |
| Data Analyst |
nous |
deepseek-v4-flash:free |
Light data queries |
Deleted (2026-05-27): council, explorer, librarian — generic boilerplate, no custom SOUL, no Discord bots.
Tier 2 — Paid Heavy Lifters (Xiaomi MiMo v2.5 Pro)
| Profile |
Reasoning |
Role |
| Coder |
high |
Implementation — builds features, fixes bugs, writes tests |
| Reviewer |
xhigh |
Code review gate — blocks unsafe/untested work before merge |
| Architect |
high |
Structural design, architecture decisions, module planning |
| Security |
high |
Security audits, vulnerability assessment |
| Debugger |
high |
Root cause analysis, systematic debugging |
Tier 3 — DeepSeek Direct (deepseek-v4-pro)
Premium reasoning via DeepSeek API. Best quality for research, analysis, and code generation.
| Profile |
Provider |
Model |
Reasoning |
Role |
| Oracle |
deepseek |
deepseek-v4-pro |
high |
Market intelligence, trading research, analysis |
| Researcher |
deepseek |
deepseek-v4-pro |
high |
Deep research, source synthesis, investigation |
| Designer |
openrouter |
deepseek/deepseek-chat-v3-0324 |
high |
UI/graphics/image generation — routed via OpenRouter (no DEEPSEEK_API_KEY set) |
Model Selection Guide
When choosing a model for a new profile, match to the work type:
| Work type |
Best model |
Why |
| Conversation / routing / triage |
mimo-v2.5-pro (xiaomi) |
Cheapest, good enough for decisions |
| Code generation (HTML/CSS, JS, shaders, templates) |
deepseek-v4-pro |
Strongest at code output |
| Deep reasoning (research, analysis, strategy) |
deepseek-v4-pro |
Best reasoning quality |
| Text / knowledge work (writing, docs, curation) |
owl-alpha (openrouter) |
Good prose, knowledge retrieval |
| Lightweight checks (smoke tests, quick lookups) |
deepseek-v4-flash:free (nous) |
Free tier, fast |
Rule of thumb: If the profile generates code (even HTML/CSS for UI), use deepseek-v4-pro. If it routes or triages, use mimo. If it writes prose, use owl-alpha.
Config templates
Xiaomi:
model:
provider: xiaomi
default: xiaomi/mimo-v2.5-pro
base_url: https://token-plan-sgp.xiaomimimo.com/v1
DeepSeek Direct:
model:
provider: deepseek
default: deepseek-v4-pro
base_url: https://api.deepseek.com
DeepSeek via OpenRouter (no DEEPSEEK_API_KEY needed):
model:
provider: openrouter
default: deepseek/deepseek-chat-v3-0324
Nous free tier:
model:
provider: nous
default: deepseek/deepseek-v4-flash:free
base_url: https://inference-api.nousresearch.com/v1
Reasoning Effort Map
| Level |
Profiles |
When to use |
none |
foreman, explorer, secretary, librarian, devops, data-analyst |
Routing, triage, lookups |
high |
coder, architect, security, debugger |
Deep reasoning — implementation, design |
high |
oracle, researcher, designer |
Deep reasoning via DeepSeek direct — research, analysis, code generation |
xhigh |
reviewer, council |
Maximum depth — gatekeeping, strategy |
Notes
- Each profile needs its own API keys in its own
.env — profiles do NOT inherit from root ~/.hermes/.env. Root .env is only read by processes that resolve to the default profile (senna). All other profiles must have the relevant key in ~/.hermes/profiles/<name>/.env.
- OpenRouter as aggregator fallback: If a profile's direct provider key is missing (e.g. DEEPSEEK_API_KEY), switch to OpenRouter with
provider/model format instead of adding a new key. Designer uses this pattern. Verify current model IDs at openrouter.ai/models before committing.
- Xiaomi credentials:
XIAOMI_API_KEY in each profile's .env. Check with grep "XIAOMI_API_KEY" ~/.hermes/profiles/<name>/.env.
- DeepSeek credentials:
DEEPSEEK_API_KEY in each profile's .env.
- OpenRouter credentials:
OPENROUTER_API_KEY in each profile's .env.
- Model name format: Always use
provider/model format (e.g. xiaomi/mimo-v2.5-pro, deepseek/deepseek-v4-pro). Bare model names like mimo-v2.5-pro may work but the provider prefix is the canonical form.
- The
:free suffix on deepseek v4 flash maps to Nous Portal's free-tier endpoint.
- User explicitly avoids deepseek-r1-0528.
- Senna stays on xiaomi/mimo — user has 1B MiMo tokens, reasonable for session profile.
- Oracle and Researcher use DeepSeek v4 Pro directly for best reasoning quality.
- Always verify with actual config:
grep -A5 "^model:" ~/.hermes/profiles/<name>/config.yaml
- See
references/profile-key-inheritance.md for the full explanation of why profiles don't inherit root .env keys.
1---
2name: profile-model-fleet
3description: Current model assignments for all Hermes profiles across multi-provider fleet (Nous + Xiaomi + DeepSeek), with pricing, reasoning_effort, and swarm role mapping. Use this to re-apply after resets or to audit profile-level model configs.
4version: 3.1.0
5platforms: [macos, linux]
6triggers: [profile model, model assignment, profile fleet, nous provider models, xiaomi provider models, deepseek provider models, profile config model, swarm organization]
7metadata:
8 source: https://skillmd.com/skills/theheavenlyd3mon/profile-model-fleet
9 hermes:
10 tags: [profile, model, nous, xiaomi, deepseek, config, pricing, fleet, swarm]
11 related: [foreman-orchestration, kanban-orchestrator, hermes-companion-apps]
12---
13
14# Profile Model Fleet — Three-Provider Assignments
15
16The fleet uses three providers:
17- **Xiaomi MiMo** (`provider: xiaomi`, `base_url: https://token-plan-sgp.xiaomimimo.com/v1`) — primary heavy lifter, 1B token giveaway
18- **DeepSeek Direct** (`provider: deepseek`, `base_url: https://api.deepseek.com`) — premium reasoning for research/market analysis
19- **Nous Portal** (`provider: nous`, `base_url: https://inference-api.nousresearch.com/v1`) — free tier for lightweight workers (deepseek-v4-flash:free)
20- **OpenRouter** (`provider: openrouter`) — used by secretary for owl-alpha
21
22## Profile → Model Map (as of 2026-05-27)
23
24### Senna (default chat session)
25| | |
26|---|---|
27| **Model** | `mimo-v2.5-pro` |
28| **Provider** | `xiaomi` |
29| **Base URL** | `https://token-plan-sgp.xiaomimimo.com/v1` |
30
31### Oracle (market intelligence)
32| | |
33|---|---|
34| **Model** | `deepseek-v4-pro` |
35| **Provider** | `deepseek` |
36| **Base URL** | `https://api.deepseek.com` |
37
38> **NOTE (2026-05-27):** Researcher was briefly switched to xiaomi/mimo-v2.5-pro due to missing `DEEPSEEK_API_KEY` in its profile `.env`. Key was copied from root `.env`, researcher switched back to deepseek-v4-pro with `reasoning_effort: high`.
39
40### Tier 1 — Lightweight (Xiaomi + OpenRouter, reasoning: none)
41Fast, cheap workers for routing, triage, light tasks.
42| Profile | Provider | Model | Role |
43|---|---|---|---|
44| **Foreman** | xiaomi | mimo-v2.5-pro | Swarm orchestrator — decomposes tasks, routes to specialists |
45| **Secretary** | openrouter | owl-alpha | Inbox triage, docs, writing |
46| **DevOps** | nous | deepseek-v4-flash:free | Ops health checks, dependency updates |
47| **Data Analyst** | nous | deepseek-v4-flash:free | Light data queries |
48
49**Deleted (2026-05-27):** council, explorer, librarian — generic boilerplate, no custom SOUL, no Discord bots.
50
51### Tier 2 — Paid Heavy Lifters (Xiaomi MiMo v2.5 Pro)
52| Profile | Reasoning | Role |
53|---|---|---|
54| **Coder** | `high` | Implementation — builds features, fixes bugs, writes tests |
55| **Reviewer** | `xhigh` | Code review gate — blocks unsafe/untested work before merge |
56| **Architect** | `high` | Structural design, architecture decisions, module planning |
57| **Security** | `high` | Security audits, vulnerability assessment |
58| **Debugger** | `high` | Root cause analysis, systematic debugging |
59
60### Tier 3 — DeepSeek Direct (deepseek-v4-pro)
61Premium reasoning via DeepSeek API. Best quality for research, analysis, and code generation.
62| Profile | Provider | Model | Reasoning | Role |
63|---|---|---|---|---|
64| **Oracle** | deepseek | deepseek-v4-pro | `high` | Market intelligence, trading research, analysis |
65| **Researcher** | deepseek | deepseek-v4-pro | `high` | Deep research, source synthesis, investigation |
66| **Designer** | openrouter | deepseek/deepseek-chat-v3-0324 | `high` | UI/graphics/image generation — routed via OpenRouter (no DEEPSEEK_API_KEY set) |
67
68### Model Selection Guide
69
70When choosing a model for a new profile, match to the work type:
71
72| Work type | Best model | Why |
73|---|---|---|
74| **Conversation / routing / triage** | mimo-v2.5-pro (xiaomi) | Cheapest, good enough for decisions |
75| **Code generation** (HTML/CSS, JS, shaders, templates) | deepseek-v4-pro | Strongest at code output |
76| **Deep reasoning** (research, analysis, strategy) | deepseek-v4-pro | Best reasoning quality |
77| **Text / knowledge work** (writing, docs, curation) | owl-alpha (openrouter) | Good prose, knowledge retrieval |
78| **Lightweight checks** (smoke tests, quick lookups) | deepseek-v4-flash:free (nous) | Free tier, fast |
79
80**Rule of thumb:** If the profile generates code (even HTML/CSS for UI), use deepseek-v4-pro. If it routes or triages, use mimo. If it writes prose, use owl-alpha.
81
82## Config templates
83
84Xiaomi:
85```yaml
86model:
87 provider: xiaomi
88 default: xiaomi/mimo-v2.5-pro
89 base_url: https://token-plan-sgp.xiaomimimo.com/v1
90```
91
92DeepSeek Direct:
93```yaml
94model:
95 provider: deepseek
96 default: deepseek-v4-pro
97 base_url: https://api.deepseek.com
98```
99
100DeepSeek via OpenRouter (no DEEPSEEK_API_KEY needed):
101```yaml
102model:
103 provider: openrouter
104 default: deepseek/deepseek-chat-v3-0324
105```
106
107Nous free tier:
108```yaml
109model:
110 provider: nous
111 default: deepseek/deepseek-v4-flash:free
112 base_url: https://inference-api.nousresearch.com/v1
113```
114
115## Reasoning Effort Map
116
117| Level | Profiles | When to use |
118|---|---|---|
119| `none` | foreman, explorer, secretary, librarian, devops, data-analyst | Routing, triage, lookups |
120| `high` | coder, architect, security, debugger | Deep reasoning — implementation, design |
121| `high` | oracle, researcher, designer | Deep reasoning via DeepSeek direct — research, analysis, code generation |
122| `xhigh` | reviewer, council | Maximum depth — gatekeeping, strategy |
123
124## Notes
125
126- **Each profile needs its own API keys in its own `.env`** — profiles do NOT inherit from root `~/.hermes/.env`. Root `.env` is only read by processes that resolve to the default profile (senna). All other profiles must have the relevant key in `~/.hermes/profiles/<name>/.env`.
127- **OpenRouter as aggregator fallback**: If a profile's direct provider key is missing (e.g. DEEPSEEK_API_KEY), switch to OpenRouter with `provider/model` format instead of adding a new key. Designer uses this pattern. Verify current model IDs at openrouter.ai/models before committing.
128- Xiaomi credentials: `XIAOMI_API_KEY` in each profile's `.env`. Check with `grep "XIAOMI_API_KEY" ~/.hermes/profiles/<name>/.env`.
129- DeepSeek credentials: `DEEPSEEK_API_KEY` in each profile's `.env`.
130- OpenRouter credentials: `OPENROUTER_API_KEY` in each profile's `.env`.
131- **Model name format**: Always use `provider/model` format (e.g. `xiaomi/mimo-v2.5-pro`, `deepseek/deepseek-v4-pro`). Bare model names like `mimo-v2.5-pro` may work but the provider prefix is the canonical form.
132- The `:free` suffix on deepseek v4 flash maps to Nous Portal's free-tier endpoint.
133- User explicitly avoids deepseek-r1-0528.
134- Senna stays on xiaomi/mimo — user has 1B MiMo tokens, reasonable for session profile.
135- Oracle and Researcher use DeepSeek v4 Pro directly for best reasoning quality.
136- Always verify with actual config: `grep -A5 "^model:" ~/.hermes/profiles/<name>/config.yaml`
137- See `references/profile-key-inheritance.md` for the full explanation of why profiles don't inherit root `.env` keys.