$ AI PriceCompare
Updated July 23, 2026 · Refreshed weekly

Seven AI chatbots,
one honest
price sheet.

Every popular AI service charges roughly $20 a month and quietly disagrees about what you get for it. We translate the marketing into a single side-by-side ledger — free tiers, paid tiers, the genuine differences — and update it the day prices change.

Read the ledger ↓ Developer API rates
Compares7services
Tracks23plans
API models19priced
Languages5translated
§ 01 — Pick by need

If you only have ten minutes,
start here.

01
For writing & everyday chat
ChatGPT
$0–200Go is $8, Plus $20
↗
02
For long documents & coding
Claude
$20Pro is the sweet spot
↗
03
For Google docs & video gen
Gemini
$19.99+ 2 TB storage
↗
04
For research with sources
Perplexity
$0–20free is generous
↗
05
For Office 365 workflows
Copilot
$20needs M365
↗
06
For free, unlimited reasoning
DeepSeek
$0100% free, no paywall
↗
07
For X integration & real-time
Grok
$30priciest mainstream
↗
§ 02 — The ledger

All seven, side-by-side.
No asterisks.

Filter
Sort
ChatGPT
by OpenAI
WritingImagesVoice
Read profile →
Free
Free
Free
  • GPT-5.4 mini
  • Limited GPT-5.6 Terra
  • Voice mode
  • File uploads
  • Image generation (limited)
Go
$8 /mo
ChatGPT Go
  • GPT-5.6 Terra + GPT-5.4 mini
  • Higher caps than Free
  • Image generation
  • Voice mode
  • Memory & projects
Plus
$20 /mo
ChatGPT Plus
  • GPT-5.6 Sol + Terra + GPT-5.4 mini
  • 20 GPT-5.6 Sol msgs / week
  • 100 GPT-5 thinking / week
  • Unlimited GPT-5.4 nano
  • 10 Deep Research / mo
Pro
$200 /mo
ChatGPT Pro
  • Unlimited GPT-5.6 Sol + 5.5 pro
  • GPT-5.5 pro mode
  • 250 Deep Research / mo
  • Priority access
  • Sora video generation
Claude
by Anthropic
WritingCodingDocuments
Read profile →
Free
Free
Free
  • Claude Sonnet 5
  • Limited daily usage
  • Projects
  • Chat history
Pro
$20 /mo
Claude Pro · $17/mo with annual ($200 upfront)
  • Fable 5 + Opus 5 + Sonnet 5
  • Fable 5 via usage credits only (since Jul 20); $100 one-time credit
  • ~45 msgs / 5 hr
  • Research access
  • Google Workspace
Max
$100–200 /mo
Claude Max
  • Opus 5 default model
  • Fable 5 included (≤50% of weekly limit, ongoing)
  • $100: 5× Pro usage
  • $200: 20× Pro usage
  • Highest limits
Gemini
by Google
WritingImagesVideo (Veo)
Read profile →
Free
Free
Free
  • Gemini 3.6 Flash
  • Multimodal input
  • Basic features
Plus
$4.99 /mo
AI Plus
  • Gemini 3 Pro
  • Gemini Omni Flash (video)
  • 2× Free usage limits
  • 400 GB storage
  • NotebookLM
Pro
$19.99 /mo
AI Pro
  • Gemini 3.1 Pro
  • Gemini Omni Flash (video)
  • 1M token context
  • Deep Research
  • Veo 3.1 Lite
Ultra
$99.99–199.99 /mo
AI Ultra
  • $99.99: 5× Pro usage
  • $199.99: 20× Pro usage
  • Gemini Omni Flash (video)
  • Veo 3.1 video (1080p)
  • Deep Think & Project Genie
Copilot
by Microsoft
OfficeExcelWord
Read profile →
Free
Free
Free
  • Basic Copilot
  • Standard features
  • Limited access
—
—
Copilot Pro (discontinued)
  • No longer available for new subscribers
  • Existing subscribers supported until Aug 1, 2026
  • Microsoft now points individuals to M365 Premium
Premium
$16.67 /mo
M365 Premium · $199.99/year
  • 6 family accounts
  • Full Office apps
  • Up to 6 TB OneDrive shared
  • Premium Copilot AI
  • Workflow automation
Perplexity
by Perplexity AI
SearchResearchCitations
Read profile →
Free
Free
Free
  • Unlimited Quick searches
  • 5 Pro searches/day
  • Standard model
  • Search history
Pro
$20 /mo
Pro · $200/year
  • 300+ Pro searches/day
  • GPT-5.6, Claude Sonnet 5, Sonar 2
  • Unlimited file uploads
  • Image generation
  • $5 API credits/mo
Max
$200 /mo
Max · $2,000/year
  • Unlimited advanced models
  • GPT-5.6 + Claude Opus 5
  • Unlimited Labs
  • Early access
  • Priority support
DeepSeek
by DeepSeek AI
FreeReasoningCoding
Read profile →
Free
Free
Free
  • DeepSeek V4-Flash + V4-Pro
  • Unlimited usage
  • Chain-of-thought
  • No ads or paywalls
  • Web, iOS, Android
—
—
No paid plan
  • Currently free only
—
—
API available
  • Pennies per million tokens
Grok
by xAI
Real-timeX integrationImage edit
Read profile →
Free
Free
Free
  • 2 prompts / 2 hr
  • Basic Grok
  • X integration
  • Real-time search
Premium+
$30–40 /mo
SuperGrok / X Premium+
  • SuperGrok $30: higher limits
  • X Premium+ $40: Grok 4.5
  • Deep search
  • Image editing
  • Real-time X
Heavy
$300 /mo
SuperGrok Heavy
  • Grok 4.5 Heavy
  • Increased limits
  • Frontier-level perf
  • Priority support

Footnote Prices in USD. Annual billing is typically 10–20% cheaper. Tap any service to read its full provider profile.

§ 03 — Provider profiles

Read the long version.

Each AI service gets its own dossier — pricing history, what it's actually good at, what it's quietly bad at, who runs it, and how it compares to the others.

§ 03b — Head-to-head

Pick two. Read the verdict.

Most people aren't choosing from seven AIs — they're choosing between two. Here are the most common matchups, with an honest call on which one to pick for your situation.

ChatGPT vs Claude vs Gemini→ ChatGPT vs Claude→ ChatGPT vs Gemini→ Claude vs Gemini→ Claude vs ChatGPT (coding)→ ChatGPT vs Perplexity→ Gemini vs Copilot→ DeepSeek vs ChatGPT→ Grok vs ChatGPT→
§ 03c — Beyond chatbots

The AI app catalog.

Chatbots are only half the story. Image generators, video tools, coding assistants, music makers — here's what the popular ones cost.

Browse all AI apps →

§ 04 — API rates

For developers, the real bill.

API pricing is per-million-tokens. Roughly: 1M tokens ≈ 750,000 words ≈ a stack of paperbacks. Output tokens cost 3–5× input tokens. Below: every model worth using right now, sorted by output price.

Provider
Model
Input / 1M
Output / 1M
Context
Mistral
Ministral 3 14B
Budget multimodal; replaced Pixtral 12B
$0.20
$0.20
256K
DeepSeek
DeepSeek V4-Flash
Fast, cheap reasoning
$0.14
$0.28
1M
Google
Gemini 2.5 Flash-Lite
Fast, affordable
$0.10
$0.40
1M
Mistral
Mistral Small 4
Mid-tier general; replaced Small 3.1
$0.15
$0.60
256K
DeepSeek
DeepSeek V4-Pro
Top reasoning, cheap
$0.43
$0.87
1M
Mistral
Codestral
Code generation
$0.30
$0.90
128K
OpenAI
GPT-5.4 nano
Fastest, cheapest
$0.20
$1.25
400K
Mistral
Mistral Large 3
Mistral flagship
$0.50
$1.50
256K
xAI
Grok Build 0.1
Code tasks; replaced Grok Code Fast 1; >200K tokens: $2/$4
$1.00
$2.00
256K
Google
Gemini 3.5 Flash-Lite
Fastest Flash tier (~350 tok/s); succeeds 3.1 Flash-Lite
$0.30
$2.50
1M
Google
Gemini 2.5 Flash
Best price / performance
$0.30
$2.50
1M
xAI
Grok 4.3
Reasoning workhorse; replaced Grok 4; >200K tokens: $2.50/$5
$1.25
$2.50
1M
Meta
Muse Spark 1.1
Meta's first paid model; agentic (Jul 2026)
$1.25
$4.25
1M
OpenAI
GPT-5.4 mini
Cheaper, well-defined tasks
$0.75
$4.50
400K
Anthropic
Claude Haiku 4.5
Fast, near Sonnet 4 quality
$1.00
$5.00
200K
OpenAI
GPT-5.6 Luna
Fast & cheapest 5.6 tier
$1.00
$6.00
1M
xAI
Grok 4.5
Opus-class flagship (Jul 2026); >200K tokens: $4/$12
$2.00
$6.00
500K
Google
Gemini 3.6 Flash
New workhorse (Jul 2026); 17% fewer output tokens, Computer Use
$1.50
$7.50
1M
Google
Gemini 2.5 Pro
Long context (1M)
$1.25
$10.00
1M
Google
Gemini 3.1 Pro
Frontier multimodal (>200K: $4/$18)
$2.00
$12.00
1M
OpenAI
GPT-5.6 Terra
Balanced 5.6; ~2× cheaper than Sol
$2.50
$15.00
1M
Anthropic
Claude Sonnet 5
New Sonnet flagship; intro $2/$10 to Aug 31
$3.00
$15.00
1M
Anthropic
Claude Opus 5
New Opus flagship (Jul 24); near-Fable 5 at half the price
$5.00
$25.00
1M
OpenAI
GPT-5.6 Sol
New flagship (Jul 2026); Ultra mode spawns sub-agents
$5.00
$30.00
1M
Anthropic
Claude Fable 5
Most powerful; restored Jul 1 after export-control pause
$10.00
$50.00
1M
OpenAI
GPT-5.5 pro
Most capable, smartest
$30.00
$180.00
1M
Worked example

Translating a 2,500-word essay (≈ 6,666 tokens, half in / half out):

  • Claude Haiku 4.5$0.020
  • GPT-5 mini$0.008
  • Gemini 2.5 Flash$0.009
  • Claude Sonnet 5$0.060
  • Claude Opus 5$0.100
  • GPT-5.4 pro$0.450

TipStart cheap, upgrade only if the output isn't good enough. The cheapest model is usually 95% as good for simple tasks.

§ 05 — Cost calculator

What will it cost you?

§ 06 — Frequently asked

Common doubts, addressed.

Start with the free tiers of ChatGPT and Gemini — both let you try the real product without a credit card. ChatGPT is the most forgiving for general tasks; Gemini is better if you live in Gmail and Docs. Most people find a $20/month plan more than sufficient once they upgrade.

Coordination, mostly. ChatGPT Plus set the anchor in early 2023 at $20, and competitors followed because customers learned to expect it. The real differences live in usage caps, model access, and adjacent perks. Two $20 plans can deliver wildly different value.

Yes — and many people do. Free plans are deliberately useful enough to keep you in the ecosystem, with the cost being slower responses, fewer messages per day, or no access to the best models. DeepSeek is the outlier with no paywall at all.

ChatGPT is best at general tasks and image generation. Claude excels at long documents, careful reasoning, and software engineering. Gemini wins on Google integration, video (Veo 3), and very long contexts.

Generally yes, but providers may use chats to train future models unless you opt out. ChatGPT and Claude both expose this setting; Gemini handles it via Google account controls. Avoid sharing passwords, financial details, or sensitive personal information regardless.

Only if you're hitting the usage caps on the $20 plans daily. Pro tiers exist for people running serious workflows. If you're a casual user, $20 is the sweet spot and $200 is mostly waste.

Yes — every consumer AI plan is monthly with no commitment. Your features remain active through the end of the billing period. Annual plans give a 10–20% discount but lock you in for the year.

Perplexity offers an Education Pro plan at $4.99/month with student verification. Other providers don't currently offer formal student discounts, though free tiers are usually enough for homework.

Methodology

How we keep this honest.

Every price is sourced directly from the provider's pricing page on the date stamped in the masthead. We do not take affiliate fees from any AI provider. When a price changes, the old number is struck through and the new one is dated.

Plan descriptions are paraphrased from official documentation, then cross-checked against independent reviews. Where providers disagree with our characterization, we publish their statement alongside ours.

Read full methodology →