Services
DatabaseSchedulerLibrarianBoard

Model Routing

Configure primary, secondary, and tertiary LLM providers per agent

Anthropic via OpenRouter

16

Google via OpenRouter

2

OpenAI direct

1

Groq direct

1

Claude Max (local)

1

Total agents

21

Agent Model Assignments

Click a model slot to change it. Changes require confirmation and emit an audit event.

#AgentRiskPrimarySecondaryTertiary
01Chief of StaffHIGH
02Product EngineerMEDIUM-HIGH
03Product ManagerMEDIUM
04Business DevMEDIUM
05Ops ManagerLOW-MEDIUM
06Code ReviewerCRITICAL
07VerifierLOW
08Sales ResearcherLOW
09Inbox ManagerLOW-MEDIUM
10DevOps EngineerCRITICAL
11Backend/ML EngineerMEDIUM-HIGH
12Best Practices BotMEDIUM
13QA TesterLOW
14InnovatorMEDIUM
15SynthesizerLOW
16ArchitectMEDIUM-HIGH
17CFOCRITICAL
20Director of EngineeringHIGH
21Director of OperationsHIGH
22Director of ReliabilityHIGH
23Inspector GeneralMEDIUM

Available Models

Direct APIBilled per token via provider API key
Local BridgeRoutes via local CLI (Claude Max / Gemini subscription) — $0 per call
Local OllamaRuns on Mac Studio GPU — $0, no quota, offline capable
Free APIProvider free tier — $0 but rate-limited

Direct API (Billed)

Claude Opus 4.7

anthropic

claude-opus-4-7

$15/$75/1M

Anthropic API direct billing. Latest Opus — best for complex reasoning, agents.

Claude Opus 4.6

anthropic

claude-opus-4-6

$15/$75/1M

Anthropic API direct billing. Previous Opus generation.

Claude Sonnet 4.6

anthropic

claude-sonnet-4-6

$3/$15/1M

Anthropic API direct billing. Best quality/cost for most tasks.

Claude Sonnet 4.5

anthropic

claude-sonnet-4-5

$3/$15/1M

Anthropic API direct billing. Previous Sonnet generation.

Claude Haiku 4.5

anthropic

claude-haiku-4-5

$0.8/$4/1M

Anthropic API direct billing. Fastest Claude model.

Claude Haiku 4.5 (Oct)

anthropic

claude-haiku-4-5-20251001

$0.8/$4/1M

Full model ID form of claude-haiku-4-5 (2025-10-01 release).

GPT-4o

openai

gpt-4o

$5/$15/1M

OpenAI API direct billing. Strong general-purpose model.

GPT-4o Mini

openai

gpt-4o-mini

$0.15/$0.6/1M

OpenAI API direct billing. Budget option for simple tasks.

Gemini 2.5 Pro

google

gemini-2.5-pro

$1.25/$5/1M

Google AI API direct billing. Strong at long context and reasoning.

Gemini 2.5 Flash Preview

google

gemini-2.5-flash-preview-04-17

$0.1/$0.4/1M

Google AI API direct billing. Fast and cost-effective.

Grok 3

xai

grok-3

$3/$15/1M

xAI API. Director of Contrarian Review. Strong at adversarial analysis.

Grok 3 Mini

xai

grok-3-mini

$0.3/$0.5/1M

xAI API. Budget Grok variant for fast tasks.

Local Bridges ($0)

Claude Max (local bridge)

local

claude-max-bridge

Free

Routes via `claude --print` using Ethan's Claude Max subscription. $0 per call, but uses personal usage quota. Max concurrency=3, p95 latency ~3-5 min under load.

Gemini CLI (local bridge)

local

gemini-bridge

Free

Routes via `gemini -p` using Google subscription. $0 per call, tight daily quota (~50 calls). Fallback only.

Local Ollama ($0)

Phi-4 14B (local)

ollama

phi4:14b

Free

⚠ Tool calls broken (registry change 2026-04-19). Use llama3.1:8b for tool-capable tasks. Runs on Mac Studio GPU, $0.

Gemma 3 27B (local)

ollama

gemma3:27b

Free

⚠ No tool call support. Text generation only. Runs on Mac Studio GPU, $0.

Llama 3.1 8B (Ollama)

ollama

llama3.1:8b

Free

Runs on Mac Studio via Ollama. $0, no quota, tool-call capable. Best local fallback for agents.

Free API Tier

Llama 3.3 70B

groq

llama-3.3-70b-versatile

Free

Groq free tier. Rate-limited (~30 req/min). No per-token cost but has daily caps.

Llama 3.1 8B Instant

groq

llama-3.1-8b-instant

Free

Groq free tier. Very fast inference, rate-limited. Best for simple/fast tasks.

Routing Change History

Last 20 model routing changes

No routing changes recorded yet.