AI Chatbot Pricing Index: April 2026
This month's index covers nine platforms. The notable development is happening at OpenAI: GPT-4o reaches end-of-life across all ChatGPT plans on April 3, completing a transition that began when the company shifted its model defaults toward GPT-5 in late 2025. Elsewhere, DeepSeek released a new flagship model, xAI added a $300/month reasoning tier, and Microsoft announced commercial pricing changes taking effect in July.
The data below reflects pricing current as of April 1, 2026.
The Data
| Chatbot | Free Tier | Mid (~$20/mo) | Premium |
|---|---|---|---|
| ChatGPT | GPT-5, voice, image gen | Go $8/mo · Plus $20/mo | Pro $200/mo |
| Claude | Sonnet 4.5, Projects | Pro $20/mo | Max 5x $100/mo · Max 20x $200/mo |
| Gemini | Limited Gemini 3 Flash | AI Pro $19.99/mo | AI Ultra $249.99/mo |
| Perplexity | 5 Pro Searches/day | Pro $20/mo ($200/year) | Enterprise $40/seat/mo |
| Grok | Basic Grok 4 | SuperGrok $30/mo | SuperGrok Heavy $300/mo |
| Le Chat (Mistral) | Full chat access | Pro $14.99/mo | API pay-as-you-go |
| Microsoft Copilot | Included with M365 | Pro $20/mo | Business $18-21/user/mo |
| GitHub Copilot | 2,000 completions/50 chats | Pro $10/mo | Pro+ $39/mo · Enterprise $39/user/mo |
| DeepSeek | Unlimited web app | API only: V4 $0.30/$0.50/1M tokens | No subscription tier |
What Changed in April
GPT-4o reaches end-of-life on April 3. OpenAI's GPT-4o, which powered ChatGPT for most of 2024 and 2025, is being retired from all plans beginning April 3, 2026. The immediate impact is for Business and Enterprise customers using custom GPTs built on GPT-4o -- those integrations will need to be migrated to GPT-5 or GPT-5.4 equivalents. Consumer Plus and Pro subscribers have been on GPT-5 as the default model since late 2025, so most users will notice no change. For custom GPT developers, the deadline is effectively now.
DeepSeek released V4. DeepSeek's latest model, V4, launched in early March and is now the default when accessing DeepSeek via web or API. The pricing increased marginally from V3's $0.28/$0.42 per million tokens to V4's $0.30/$0.50 -- a roughly 10-15% increase that accompanies a significant capability jump. On SWE-bench Verified, V4 scores 81% versus V3's 69%, putting it alongside GPT-5 and Claude Sonnet 4.5 on coding benchmarks. Cache hit pricing at $0.03 per million input tokens makes V4 one of the most cost-effective options for applications with high prompt reuse. The free web app continues to include V4 at no cost with daily usage limits.
xAI added a $300/month SuperGrok Heavy tier. Grok's lineup now has a new ceiling: SuperGrok Heavy at $300/month, designed for users who need extended reasoning via Grok 4 Heavy, xAI's most computationally intensive model. This sits above the standard SuperGrok ($30/month) by a factor of ten. For context, it is priced higher than Claude Max 20x ($200) and ChatGPT Pro ($200), making it the highest-priced individual AI subscription currently on the market. The tier targets users running complex multi-step reasoning tasks where Grok's real-time X integration adds meaningful value.
Microsoft announced July pricing changes. Microsoft has confirmed that commercial Microsoft 365 pricing will update on July 1, 2026. The company is adding expanded AI, security, and management capabilities to base subscriptions ahead of that change. For Copilot Business subscribers, the current promotional rate of $18/user/month (available through June 30) is relevant context for planning purposes. Microsoft has not released the specific post-July rates, but customers renewing in the next quarter should factor this in.
What Did Not Change
The $20/month tier held across the major platforms. ChatGPT Plus, Claude Pro, Gemini AI Pro, and Perplexity Pro all remain at $19.99-$20/month -- a price point that has been stable since mid-2025.
DeepSeek's consumer pricing remains zero. There are no subscription tiers for web or mobile access; the free app is the primary access point for most users. API pricing for developers increased slightly with the V4 launch but remains by far the lowest among frontier-class models.
Mistral Le Chat's $14.99/month Pro plan remains the lowest-priced premium subscription among the major platforms. It includes access to Mistral's current models, document uploads, image generation, and a no-training-data policy for paid subscribers.
Notes on Use Case Fit
Coding: GitHub Copilot Pro at $10/month remains the baseline for AI-assisted code completion. DeepSeek V4's API is worth evaluating for custom development tooling where latency and cost matter -- $0.30/$0.50 per million tokens with 81% SWE-bench performance is the strongest cost-per-capability ratio in the market right now.
Research: Perplexity Pro's core value proposition remains intact. The deep research limit issues from March have been partially addressed. The $200/year annual plan continues to offer better value than monthly billing if Perplexity is a primary research workflow.
Long documents: Claude's 1M-token context window (in beta for Max subscribers) and Gemini AI Pro's standard 1M-token context remain the practical choices for large-file analysis. Claude Sonnet 4.5 at the $20/month Pro tier handles most document tasks within its standard context window.
General use: Free tiers from ChatGPT, Claude, Gemini, and DeepSeek are more capable today than they were a year ago. Most users will not hit a meaningful functional wall on the free tier. Consistently hitting usage limits is the clearest signal to upgrade to a paid plan.
Full profiles and feature ratings for each platform in this index are at chatbot.gallery. For recommendations organized by task rather than price, see About.chat's guide to the best AI chatbots.
Pricing verified against official sources as of April 1, 2026. This index is updated monthly.