- OpenAI Dots cost: Your first Dot is included with ChatGPT Pro and Business Premium — no extra fee.
- Chat is free, work is not: Chatting with a Dot does not touch your limits; real tasks run through Codex usage.
- Pro plan change: The $200 Pro plan reopened with roughly half the Codex usage it previously included.
- Ultrafast speed: Available only on the new $500/month Pro 500 plan for ChatGPT users.
- Regional limits: Dots are not yet available in Europe, the UK, or Switzerland, and are adults-only on Pro.
OpenAI Dots Cost Explained: What You Actually Pay For
OpenAI Dots cost less than most people expect at first glance — the agent itself ships as an included feature rather than a separate product. Announced at DevDay 2026, Dots are persistent AI agents that hold ongoing jobs (monitoring bug reports, chasing vendors, drafting PRs) instead of completing one-off tasks. Each Dot runs on its own cloud computer and connects to more than 4,000 apps, including Slack, calendar, and email.
The cost structure breaks into three layers: the subscription that grants access, the Codex usage that powers real work, and optional upgrades like Ultrafast speed. Understanding how these layers interact is the key to avoiding surprise expenses.
Video Highlights:
- Dots launched at DevDay 2026 alongside more than 20 other announcements
- First Dot is included on ChatGPT Pro and Business Premium; Enterprise is in beta
- Chatting with your Dot does not consume your usage limits
- Real delegated work runs in Codex and does consume usage
- Ultrafast speed requires the new $500/month Pro 500 plan
The simple version: you pay for the plan, and the plan pays for the Dot. What you monitor is Codex usage — that is where delegated work draws down your allowance.
Plan-by-Plan Cost Comparison
Dots are live on ChatGPT Pro and Business Premium, with Enterprise in beta as of September 2026. Here is how the access tiers compare:
| Plan | Monthly Price | Dot Access | Notes |
|---|---|---|---|
| ChatGPT Free/Plus | — | Not available | Try GPT-6.1 Soul in Codex instead |
| Pro | $200/month | Included (first Dot) | Codex usage roughly halved vs. old plan; adults only |
| Business Premium | Custom | Included | Team-oriented; adds ChatGPT Space collaboration |
| Pro 500 | $500/month | Included | Adds Ultrafast speed in ChatGPT (up to 8x faster) |
| Enterprise | Custom | Beta | Shared specialist Dots (accounting, legal, support) |
The $200 Pro plan reopened with about half the Codex usage it used to include. Existing subscribers kept their old limits for roughly one month before the reduction applied — verify your current allowance before assigning your Dot heavy workloads.
Beyond the base subscription, two model-level costs matter for anyone running Dots through the API or Codex:
| Item | Input Cost | Output Cost | Best For |
|---|---|---|---|
| GPT-6.1 Soul | $2 per 1M tokens | $10 per 1M tokens | Ties flagship on coding tests at a fifth of the price |
| GPT-6 Astra | $10 per 1M tokens | $50 per 1M tokens | Maximum quality for complex tasks |
| Ultrafast API tier | 6x normal rate | 6x normal rate | Latency-sensitive apps |
Budget Setup
- Pro plan at $200/month
- GPT-6.1 Soul for Dot work
- Watch usage after plan change
Speed Setup
- Pro 500 at $500/month
- Ultrafast processing
- Best for time-critical jobs
Team Setup
- Business Premium
- Shared Spaces with tagged Dots
- Specialist Dots for accounting and legal
Step-by-Step: Estimate Your Dots Cost Before Committing
Follow this process to project your monthly spend instead of discovering it on your invoice.
Confirm Your Access Tier
Verify your plan includes Dots — Pro and Business Premium today, Enterprise via beta. Check regional availability: Pro users in Europe, the UK, and Switzerland do not have access yet.
Start With One Small Job
Give your first Dot a single low-stakes responsibility, such as watching one inbox or monitoring one Slack channel. This calibrates how much Codex usage a typical task consumes.
Measure Usage Over a Week
Track how delegated work draws down your Codex allowance. Multiply by your expected job count. Note that chatting with the Dot itself does not count against limits.
Pick the Right Model
Default to GPT-6.1 Soul, which ties the flagship Astra on OpenAI's coding benchmark (deep SWE) at roughly one-fifth the cost. Reserve Astra for tasks where quality gaps appear.
Decide on Speed Upgrades
Only move to Pro 500 ($500/month) or the 6x API rate if latency genuinely costs you money. Same prompt, same task — Ultrafast finishes up to 8x faster, but the price multiplies.
Early access reviewers report Dots can stall mid-demo and remain buggy. Budget for human oversight time as part of your true cost — an agent that needs supervision is not fully hands-off yet.
Hidden Costs and Money-Saving Tactics
The sticker price rarely tells the whole story. These factors shape your real monthly Dots spend:
| Hidden Factor | Cost Impact | Mitigation |
|---|---|---|
| Codex usage cap cut | Effectively doubles cost per unit of work on $200 Pro | Audit usage; consider Pro 500 if you hit caps |
| Sensitive-action checkpoints | Dots pause and ask before sensitive steps — time cost | Pre-approve routine job categories |
| Agent errors | Re-work and review time | Start with low-risk jobs; verify outputs |
| Ultrafast API rate | 6x normal pricing per call | Use only for latency-critical paths |
| Sign-in with ChatGPT (for builders) | Users' plans pay their own AI usage | Offloads per-user costs from your app |
If you build apps, Sign-in with ChatGPT shifts AI costs to each user's own plan — 16 apps including Devin, Notion, and Vercel launched with it. For solo builders, this removes the burden of eating the AI bill to acquire users.
Your Dots Cost Control Checklist
Before and after adopting Dots, run through this list to keep spending predictable:
Cost Control Milestones:
- Confirm Dot availability in your region and plan tier
- Assign one small pilot job to measure Codex usage
- Review the halved Pro plan usage limits before scaling up
- Set GPT-6.1 Soul as the default model for Dot tasks
- Decide whether Ultrafast speed justifies the price premium
FAQ
Q: How much do OpenAI Dots cost on their own?
Dots are not sold separately. Your first Dot is included with ChatGPT Pro ($200/month) and Business Premium, with Enterprise available in beta. The real ongoing cost is the Codex usage that powers the delegated work.
Q: Does chatting with my Dot use up my usage limits?
No. Conversations with your Dot do not touch your limits. Only real work the Dot performs runs through Codex, which does draw down your allowance.
Q: Why are Pro plan users upset about cost?
OpenAI reopened the $200 Pro plan with roughly half the Codex usage it previously included. Current subscribers kept old limits for about a month before the reduction took effect, effectively raising the cost per unit of agent work.
Q: What is the cheapest way to run Dots efficiently?
Use GPT-6.1 Soul as your workhorse model. It costs about a fifth of the flagship GPT-6 Astra ($2/$10 vs. $10/$50 per million tokens) while tying Astra on OpenAI's own coding benchmark, making it the best value for most Dot jobs.
Q: Is the $500 Pro 500 plan worth it?
Only if speed matters to you. Pro 500 includes Ultrafast processing — up to 8x faster in Codex — but costs $250 more per month than standard Pro. For background jobs that run while you sleep, standard speed is usually sufficient.