Poe

1. Should you try Poe? Overview

What it is	Poe — Quora’s Multi-Model AI Switchboard (200+ Models, Compute-Points Meter) Poe (poe.com, Quora Inc., SF/NYC, launched Feb 2023 by Adam D’Angelo) active 2026-08, no discontinuation.
Country / HQ	San Francisco, USA (Quora)
Best for	Power users bouncing Claude (code) + GPT (prose) + Gemini (summary) daily
Check before buying	Watch out: Single-model loyalists (ChatGPT Plus / Claude Pro gives better limits+memory)
Editor rating	4.5 / 5
Pricing signal	Free tier + paid from $4.99/mo

2. Features What it does

Feature	What it does
Model access	200+ text/img/video/audio bots, frontier + open-weights, per-bot “Rates” button shows point cost — New models land days–weeks after vendor launch
Multi-bot compare	Side-by-side grid, same prompt → N models, thread keeps context — Core differentiator
Custom bots	System prompt + pinned model + greeting + (formerly KB, retired May 2026) + per-message price — 1M+ community bots, creator payout ~tens of M/yr run-rate
Group Chat	200 participants, 200 models, admin controls access, each user spends own points — Nov 2025, no rival equivalent
Poe Apps / Script Bots	No-code app canvas (Feb 2025), Python model-chain scripts (late 2025) — Web-only builder, steep curve for non-trivial apps
API	OpenAI-compatible key, tool calling (OpenAI/Anthropic/Google), 500 req/min — Devs route 100+ models w/o vendor keys
UX	Web + iOS + Android + Win/mac native, synced history, file upload (PDF/img), per-message budget cap, auto-compress context, cache discounts (50% OpenAI / 90% Anthropic on reuse) — Opt-in Memory (updates daily, skips sensitive)

3. Pricing Pricing (2026-10, USD)

Plan	Price / mo	Key limits & who it fits
Free	$0	$0 ~300/day (was ~3,000 pre-Mar 2026, cut 90% unannounced) Light models only (Haiku/4o-mini), ~15–30 short msgs, no frontier, no video
Starter	$4.99	$49.99 ($4.17) 10,000/day Premium models inside daily budget, faster queue
Standard (headline)	$19.99	$199.99 ($16.67) 660,000/mo (was 1M till Jul 2026) Full GPT/Claude/Gemini + image/video, bot creation, file upload
Pro	$49.99	$499.99 ($41.67) 1.65M/mo Frontier + reasoning headroom, heavier media
Premium	$99.99	$999.99 ($83.33) 3.3M/mo Heavy multi-model + video gen
Max / Enterprise	$249.99	$2,499.99 ($208.33) 8.25M/mo Sustained frontier + team automation
Add-on points	$30 / 1M pts	1-yr expiry additive Top-up when allowance burns

4. Integration Access & integrations

Access type	Details
Access	access_type: web-app-plus-ios-plus-android-plus-win-mac-desktop-plus-api access_display: 🌐 poe.com · 📱 iOS/Android (native, rated pleasant) · 🖥️ Win/mac desktop (native) · 🔌 Poe API (api.poe.com, OpenAI-compatible, tool calling) · no Linux native, no

5. Reviews & Sources External reviews

Source	Rating	What it says
ToolRadar aggregate	4.5/5 (15 rev) — “multiple models + fair pricing + custom bots + active dev; con: message limits, Quora-owned, better direct access for power features”
AIUnpacking	“Group Chat + API + Poe Apps make it more than a router; Knowledge Base retirement May 2026 with 1-month notice + no migration tooling damaged creator trust”
AIToolTier	7.5/10 — “one sub cheaper than ChatGPT+Claude+Gemini separately if you use all three; free tier restrictive (dies in 3 convos), always one hop behind native model drops, custom bots weaker than GPTs/Claude Projects”
CompareBestAI	7.5/10 — “Premium $19.99 sweet spot for GPT-5.5/Opus/Gemini; mental overhead of watching points erodes flow; enterprise compliance details uncertain”
Zapier	“lets models talk to each other in-thread (GPT prompt → Luma image → Runway video); creator revenue share is the hook devs/distributors like”

6. Best for / Not for Fit

✔ Suitable	✘ Not suitable
• Power users bouncing Claude (code) + GPT (prose) + Gemini (summary) daily • Researchers comparing model answers on same prompt • Indie devs wanting 100+ models behind 1 OpenAI-compatible key • Bot creators wanting distribution + USD payout • Study groups / remote teams (Group Chat 200) • Consultants serving multi-model client needs	• Single-model loyalists (ChatGPT Plus / Claude Pro gives better limits+memory) • Enterprise needing SSO/SCIM/zero-retention/DPA (native vendors safer) • RAG-bot builders (KB retired May 2026, use TypingMind/custom vector) • Heavy video gen (5k–20k pts/clip, Max tier melts fast) • Offline/linux-first users (no native build, no offline) • People who hate watching a meter (points anxiety)

7. Competitors How it compares

Competitor	How it differs	Best for
ChatGPT Plus / Claude Pro / Gemini Advanced	Single-vendor depth, native memory/Codex/Projects	$20/mo each
OpenRouter	Dev-first 500+ model API gateway, no consumer chat/bots	Pay-per-token
Perplexity	Cited real-time research, not model router	$0 / $20-mo
TypingMind	Multi-model front-end + own keys + RAG upload	$59 once
Character.AI	Roleplay/companion bots, no frontier workbench	Free / $9.99-mo
Vercel AI SDK / LiteLLM	Code-layer model routing, no UX/marketplace	Free/OSS
You.com	Search-chat hybrid, fewer models	Free / $20-mo

8. Summary Verdict

Who should pick it	Power users bouncing Claude (code) + GPT (prose) + Gemini (summary) daily
Who should skip it	Single-model loyalists (ChatGPT Plus / Claude Pro gives better limits+memory)