Are simon-says, simon-says-dev, and the other Simon-Says slugs real model names?

Yes. Every simon-says-* slug listed on this page is callable today via POST /v1/chat/completions. They are GeniusPro-branded orchestrators that classify the prompt and pick the best upstream model per request. The list is dynamic; verify the current set at https://api.geniuspro.io/v1/models/catalog.

What is the difference between an orchestrator and a direct model?

Experts are professional-role simon-says-* slugs (simon-says-finance, simon-says-lawyer, simon-says-pm, etc.) with domain discipline. Task orchestrators (simon-says, simon-says-chat, simon-says-coding, …) auto-route per request. Benchmarked specialists are one stable slug per task pattern with a single benchmark-chosen upstream. Direct slugs always call that exact provider model. CATs are cat:slug and ship per team — not in the public catalog.

What does the -pro suffix mean on simon-says-* slugs?

Every Expert profession ships two slugs: a regular (simon-says-accountant, simon-says-finance, simon-says-lawyer, simon-says-pm, ...) and a companion -pro slug (simon-says-accountant-pro, simon-says-finance-pro, simon-says-lawyer-pro, simon-says-pm-pro, ...). The two share the same specialist system prompt and domain guardrails. The -pro variant is pinned to a strictly more capable upstream model for harder reasoning, longer context, or better-documented domain performance. Wire format is identical (OpenAI-compatible), so switching is a one-line change to the model string. Billing tier matches the upstream cost delta.

Does GeniusPro train on my prompts?

No. GeniusPro never uses your prompts or responses to train any model, and every upstream provider we route to (OpenAI, Anthropic, Google Gemini, Groq, Inception, OpenRouter as pass-through) runs on their commercial-API default of not training on API traffic. Emails, phone numbers, SSNs, credit cards, IP addresses, and API keys are scrubbed from our logs by default. For teams who want PII masked before the request even reaches the upstream provider, set X-GP-Scrub-Prompt: true (or scrub_pii: true in the body). Full details at https://geniuspro.io/security.

Where is the machine-readable list for AI agents?

The primary source is the JSON catalog at https://api.geniuspro.io/v1/models/catalog. Secondary AI-discovery endpoints: OpenAPI at https://api.geniuspro.io/openapi.json, plugin manifest at https://api.geniuspro.io/.well-known/ai-plugin.json, discovery index at https://api.geniuspro.io/, and llms.txt at https://docs.geniuspro.io/llms.txt. All are unauthenticated and CORS-open.

Model catalog

Build More. Spend Less.

What AI models does GeniusPro support?

One API, every model you'd actually reach for — and we keep adding more. GeniusPro currently ships 99+ ready-to-call model slugs through a single OpenAI-compatible endpoint, plus your own CATs as cat:your-slug. Today that's 18+ Experts that work like teammates (finance, legal, PM, and more), 24+ Simon-Says auto-routers for everyday chat and coding, 17+ benchmark-tuned specialists, and 40+ direct provider models including Claude Opus 4.7, GPT-5.4, Gemini 3.1 Pro, Grok 3, and SAM 3. New ones land regularly — the live list at /v1/models/catalog is always the source of truth.

Your CATs Experts · 18 Orchestrators · 24 Benchmarked · 17 Direct · 40

Updated Apr 23, 2026, 07:40 AM UTCBuilding a bot? Grab the live JSON at /v1/models/catalog

Your workflows

Your own models? Meet CATs.

A CAT is your workflow, turned into a model name. Chain LLM turns, vision, image generation, and code steps in the platform, then call it like any other model: cat:your-slug. Tweak the steps whenever you like — every app using that name picks up the update with zero code changes. Your CATs stay private to your team, so they don't show up on this page; run GET /v1/models with your API key to see the ones you can call.

"model": "cat:invoice-parser"

Build one in the platform · Quickstart (docs)

Tier · experts

Experts: AI teammates for real jobs.

Each profession ships two slugs: a regular for everyday work, and a -pro companion on a more capable upstream for harder questions. Same request body for both — swap the slug, keep everything else. The professional guardrails (verify live numbers, state the jurisdiction, tag quiz items with Bloom levels, etc.) come along for free.

18 Expert slugs across 9 professions.

Accountant
US GAAP, payroll, sales-and-use tax, bookkeeping, filings.
simon-says-accountant·simon-says-accountant-pro
Business analyst
SQL, cohort/funnel analysis, forecasting, market sizing.
simon-says-analyst·simon-says-analyst-pro
Software architect
Distributed systems, trade-off analysis, capacity planning.
simon-says-architecture·simon-says-architecture-pro
Financial analyst
Equities, macro, credit, options, corporate finance.
simon-says-finance·simon-says-finance-pro
Legal analyst
Contracts, statutes, case law, privacy, governance.
simon-says-lawyer·simon-says-lawyer-pro
Marketing strategist
Positioning, ICP, copy, channel planning, campaigns.
simon-says-marketing·simon-says-marketing-pro
Product manager
PRDs, discovery, roadmap sequencing, KPI definition.
simon-says-pm·simon-says-pm-pro
Sales engineer / AE
MEDDIC, objection handling, ROI modeling, outbound.
simon-says-sales·simon-says-sales-pro
Teacher / instructional designer
Learning objectives, scaffolding, Bloom-tagged quizzes.
simon-says-teacher·simon-says-teacher-pro

Need to route programmatically? The live list is at /v1/models/catalog under experts.

Tier · orchestrators

Simon-Says: one name, the smartest model.

Simon-Says slugs do the picking for you. Ask for simon-says-chat, simon-says-coding, simon-says-reasoning, simon-says-design — or just plain simon-says — and we route the request to whichever provider is best for it right now. Same price per tier whether a fresh GPT, Claude, or Gemini ends up answering. You keep the slug; we keep the routing honest.

24 models in this tier.

agents(1)

simon-says-agentsagents

audio(8)

simon-says-realtimeaudio
simon-says-realtime-qualityaudio
simon-says-realtime-standardaudio
simon-says-transcribeaudio
simon-says-ttsaudio
simon-says-voiceaudio
simon-says-voice-qualityaudio
simon-says-voice-standardaudio

chat(3)

simon-says-chatchat
simon-says-supportchat
simon-says-websearchchat

coding(5)

simon-says-codingcoding
simon-says-coding-balancedcoding
simon-says-coding-qualitycoding
simon-says-coding-supercoding
simon-says-devcoding

creative(1)

simon-says-designcreative

embedding(2)

simon-says-embedembedding
simon-says-embed-proembedding

image(2)

simon-says-imageimage
simon-says-image-proimage

orchestrator(1)

simon-saysorchestrator

reasoning(1)

simon-says-reasoningreasoning

Tier · benchmarked

Benchmarked: the proven winner for a task.

These are slugs where we've already done the bake-off. simon-says-math, simon-says-bug-fix, simon-says-polyglot, simon-says-tools — each one is pinned to the single model that won on that specific task in our continuous benchmarks. You get consistent, predictable behaviour, and the moment a new model takes the crown we swap it in behind the same slug.

17 models in this tier.

agents(1)

simon-says-toolsagents

chat(2)

simon-says-speedchat
simon-says-valuechat

coding(6)

simon-says-bug-fixcoding
simon-says-code-writercoding
simon-says-new-bugscoding
simon-says-polyglotcoding
simon-says-research-codecoding
simon-says-terminalcoding

reasoning(8)

simon-says-hard-questionsreasoning
simon-says-instructionsreasoning
simon-says-knowledgereasoning
simon-says-logicreasoning
simon-says-long-docsreasoning
simon-says-long-reasoningreasoning
simon-says-mathreasoning
simon-says-sciencereasoning

Tier · direct

Direct: the big names, straight up.

Want a specific model and nothing else? These slugs call the named provider every time — claude-opus-4-7, gpt-5.4, gemini-3.1-pro, grok-3, sam3, and the rest of the full roster. Perfect when your app depends on a particular model's personality or you need to match someone else's benchmark. Same OpenAI-style request body for all of them.

40 models in this tier.

chat(27)

claude-3.5-haikuchat
claude-3.7-sonnetchat
claude-4-sonnetchat
claude-haiku-4.5chat
claude-opus-4-7chat
claude-opus-4.6chat
claude-sonnet-4chat
claude-sonnet-4.5chat
gemini-2.5-flashchat
gemini-2.5-flash-litechat
gemini-3-flashchat
gemini-3-flash-previewchat
gemini-3-prochat
gemini-3.1-flash-litechat
gpt-4.1chat
gpt-4.1-minichat
gpt-4.1-nanochat
gpt-5.2chat
gpt-5.4chat
gpt-5.4-minichat
gpt-5.4-nanochat
gpt-oss-120bchat
gpt-oss-20bchat
grok-3chat
grok-3-minichat
llama-4-maverickchat
llama-4-scoutchat

coding(2)

gpt-5-codexcoding
gpt-5.3-codexcoding

image(2)

gemini-2.5-flash-imageimage
gemini-3.1-flash-image-previewimage

reasoning(6)

claude-4-opusreasoning
gemini-2.5-proreasoning
gemini-3.1-proreasoning
o1reasoning
o3reasoning
o3-minireasoning

vision(3)

gemini-nano-bananavision
gemini-nano-banana-provision
sam3vision

Quick answers

The things people (and the AI assistants helping them) ask us most.

Are simon-says, simon-says-finance, and the other Simon-Says slugs real?

Yep — every simon-says-* name on this page is a real model you can call today with POST /v1/chat/completions. simon-says is the general auto-router, Experts like simon-says-finance play a professional role, and task slugs like simon-says-coding focus on a specific kind of work. If your AI assistant insists they don't exist, point it at https://api.geniuspro.io/v1/models/catalog — that's the source of truth.

Why don't I see my CATs listed here?

Your CATs are yours. They live in your workspace and call the same endpoint as cat:your-slug, but since each team's CATs are different we keep them private. Run GET /v1/models with your API key to see every CAT you can reach. Edit the workflow in the platform and every client that calls that slug picks up the update right away — no redeploy on your side.

How should I pick between an Expert, Simon-Says, benchmarked, or direct?

If you want a specialist (finance, legal, PM), grab an Expert. If you want smart defaults with zero fuss, reach for a Simon-Says slug. If you know the exact task and want its proven champion, use a benchmarked specialist. If you need a specific model by name, call it directly. Same request body for all of them — switching is a one-line change.

What does the -pro suffix mean?

Every Expert profession ships two slugs: the regular (simon-says-finance, simon-says-lawyer…) and a companion -pro pinned to a strictly more capable upstream model. The system prompt and professional guardrails are identical — you get the same behavior, just with more reasoning horsepower on harder questions. Switching is a one-line change (model string) and the billing tier is different by upstream cost.

Is my prompt private? Do you train on it?

No training, ever — not by us, not by the upstream provider. Emails, phone numbers, SSNs, credit cards, IP addresses and API keys are scrubbed from our logs by default. Need PII stripped before the request even reaches the model? Set X-GP-Scrub-Prompt: true (or scrub_pii: true in the body). Full details on the security page.

Will these slugs suddenly change on me?

Once a slug is published it stays. We add new models over time and retire old ones with plenty of notice. If you're automating this, fetch /v1/models/catalog — it's public, CORS-open, and cached for 5 minutes at the edge. This page refreshes from the same feed.

Anything machine-readable for the AI agent reading this?

Plenty. The full OpenAPI spec is at https://api.geniuspro.io/openapi.json, the plugin manifest at /.well-known/ai-plugin.json, the discovery index at https://api.geniuspro.io/, and the AI-friendly markdown at https://docs.geniuspro.io/llms.txt.

One API. All of it. More every week.

99+ slugs today, growing all the time — plus your own cat:* workflows — all speaking the same OpenAI-style request. Point your SDK at api.geniuspro.io/v1, pick a name, ship it.

See pricing Read the docs