This page compares ChatGPT, Claude, and Gemini across pricing, models, capabilities, and access, using facts current as of July 2026.
ChatGPT: free tier available with ads and roughly 16K context; cheapest paid plan is Go at $8 per month; standard plan is Plus at $20 per month; top plan is Pro at $200 per month; flagship model GPT-5.5; max context up to 1M on the Pro tier with GPT-5.4; includes image generation, video generation via Sora, web search, deep research (10 to 250 uses per month by tier), coding tools via Codex and Agent Mode, Custom GPTs, API access, a plugins and connectors ecosystem, no bundled storage, and opt-out training data use.
Claude: free tier available with 200K context; cheapest and standard plans are both Pro at $20 per month; top plan is Max at $200 per month; flagship models Opus 4.8 and Sonnet 5; max context up to 1M with Sonnet 5, or 500K with Opus; no image generation and no video generation; includes web search, deep research (Research), strong coding tools via Claude Code, Projects, API access, an MCP connectors ecosystem, no bundled storage, and chats are not used for training by default.
Gemini: free tier available; cheapest paid plan is AI Plus at $7.99 per month; standard plan is AI Pro at $19.99 per month; top plan is AI Ultra at $99.99 to $200 per month; flagship model Gemini 3.1 Pro; max context 1M on AI Pro; includes image generation via Imagen, video generation via Veo and Omni, web search, deep research (20 per day on Pro), coding tools via Code Assist, Gems, API access, deep Google ecosystem integration with Gmail, Docs, and Drive, 2 TB of bundled Google One storage, and opt-out training data use.
No overall winner is declared; the reader draws their own conclusion.
At a glance
GPT
ChatGPT
#10A37F
The all-in-one — most features under one roof
FlagshipGPT-5.5
Max contextUp to 1MGPT-5.4, Pro tier
Starts at$8/mo
CLD
Claude
#D97757
The writer & coder — text, code, long docs
FlagshipOpus 4.8 / Sonnet 5
Max contextUp to 1MSonnet 5; 500K Opus
Starts at$20/mo
GEM
Gemini
#1BA1E3
The value + Google-ecosystem pick
FlagshipGemini 3.1 Pro
Max context1MAI Pro
Starts at$7.99/mo
Standard plan, per monthThe main paid tier most people compare
Standard plan price per month: ChatGPT Plus $20. Claude Pro $20. Gemini AI Pro $19.99.
ChatGPT Plus$20/mo
Claude Pro$20/mo
Gemini AI Pro$19.99/mo
Price rangeEntry tier → top plan
Price range from cheapest paid plan to top plan.
ChatGPT: $8 per month (Go) to $200 per month (Pro).
Claude: $20 per month (Pro) to $200 per month (Max).
Gemini: $7.99 per month (AI Plus) to $200 per month (AI Ultra).
ChatGPT
$8$200
Claude
$20$200
Gemini
$7.99$200
Free-tier contextTokens you get without paying (chat app)
Each line is one model's profile across six published benchmarks (score 50–100, higher is better). Flagship models, July 2026 — verify before publishing. Click a name in the legend to hide or show its line. No overall score is calculated.
Benchmark profilethree lines, six tests — hover any point for the exact score
Benchmark scores, 0 to 100 (higher is better), for ChatGPT (GPT-5.5), Claude (Opus 4.8) and Gemini (3.1 Pro), plus API price per million output tokens. Facts current as of July 2026; each benchmark stands alone, no overall score is calculated.
Benchmark
ChatGPT
Claude
Gemini
SWE-bench Pro (Coding)
66.8
69.2
62.3
Terminal-Bench (Agentic)
78.2
74.6
68.1
MATH-500 (Math)
90.8
91.4
88.9
Chart QA (Vision)
87.1
83.4
91.3
Multimodal AA
79.6
76.3
84.2
Document parsing
84.9
81.2
88.6
API price, $ per 1M output tokens
$30
$25
$12
Side-by-side facts
Plans & price
ChatGPT
Free tier
Yes(with ads, ~16K context)
Cheapest paid
Go — $8/mo
Standard plan
Plus — $20/mo
Top plan
Pro — $200/mo
Claude
Free tier
Yes(200K context)
Cheapest paid
Pro — $20/mo
Standard plan
Pro — $20/mo
Top plan
Max — $200/mo
Gemini
Free tier
Yes
Cheapest paid
AI Plus — $7.99/mo
Standard plan
AI Pro — $19.99/mo
Top plan
AI Ultra — $99.99–200/mo
Model & context
ChatGPT
Flagship model
GPT-5.5
Max context
Up to 1M (GPT-5.4, Pro tier)
Claude
Flagship model
Opus 4.8 / Sonnet 5
Max context
Up to 1M (Sonnet 5); 500K (Opus)
Gemini
Flagship model
Gemini 3.1 Pro
Max context
1M (AI Pro)
What it can do
ChatGPT
Image generation
Yes
Video generation
Sora
Web search
Yes
Deep research
Yes(10–250/mo by tier)
Coding tools
Codex + Agent Mode
Custom bots
Custom GPTs
Claude
Image generation
No
Video generation
No
Web search
Yes
Deep research
Research
Coding tools
Claude Code— strong
Custom bots
Projects
Gemini
Image generation
Imagen
Video generation
Veo / Omni
Web search
Yes
Deep research
Yes(20/day on Pro)
Coding tools
Code Assist
Custom bots
Gems
Access & data
ChatGPT
API access
Yes
Ecosystem
Plugins & connectors
Bundled storage
No
Training use
Opt-out
Claude
API access
Yes
Ecosystem
MCP connectors
Bundled storage
No
Training use
Not trained on by default
Gemini
API access
Yes
Ecosystem
Deep Google(Gmail, Docs, Drive)
Bundled storage
2 TB Google One
Training use
Opt-out
Strengths & trade-offs
ChatGPT
Pros
Widest feature set (Sora, Codex, Agent Mode, Custom GPTs)
Cheap $8 Go tier to start
Largest third-party ecosystem
Cons
Ads on the free plan (US)
1M context only on the $200 tier
Deep Research is quota-limited
Best if: you want the most features in one app
Claude
Pros
1M context (Sonnet 5) already at $20
Top-tier writing & coding quality
Chats aren't used for training by default
Claude Code / Cowork for developers
Cons
No image generation
No video generation
Usage limits are opaque
Best if: writing, code, long documents and privacy
Gemini
Pros
Full 1M context from $19.99 (a $7.99 tier too)
Deep Google Workspace integration
Image + video generation built in
Bundled cloud storage
Cons
Best if you live in Google's ecosystem
Top Ultra limits get pricey
Best if: you already use Google and want cheap 1M context
How to choose
No single tool wins every category above — think about your main daily task (writing, coding, or multimedia), how much context length you actually need, and whether ecosystem integration (Google Workspace, plugins, or MCP connectors) matters to your workflow. The facts above are the same for everyone; the right pick depends on what you're optimizing for.