Skip to main content

Claude vs Gemini: Opus 5.5 against Gemini 3.8 Flash

Anthropic sells four current models: Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. Google announced Gemini 4 Argon on 30 September 2026 but has released it only to cyber defenders in its Fairwind Program, so the main Gemini models you can buy today are Gemini 3.8 Flash, 3.5 Flash-Lite and 3.1 Pro, which is still a preview.

Reviewed against 30 sources. Prices, models, policies and terms change, so check the linked pages before you buy, build, publish or trade.

The short answer

For writing, code and agents, Claude is the stronger buy today: its top models outscore every Gemini model on general sale on both independent rankings read for this page. Gemini is the better buy for video and audio, generated media, a free API tier and a $4.99 chat plan.

Choose Claude if

  • You want the strongest model on general sale. Artificial Analysis scores Claude Opus 5.5 at 58 and Gemini 3.8 Flash at 41.
  • You run a high-volume job on short prompts. Haiku 5.5 is $0.10 input and $0.50 output per million tokens; Gemini 3.5 Flash-Lite is $0.30 and $2.50.
  • You need long replies. Claude writes up to 128K tokens in one response; the Gemini models on sale stop at 65,536.

Choose Gemini if

  • You send video or audio. Gemini models accept both; Anthropic lists text and image input for Claude.
  • You need images, video, music or speech generated by the same vendor. Claude models return text.
  • You want to start at $0. The Gemini API has a free tier on 3.8 Flash and 3.5 Flash-Lite, and Google AI Plus is $4.99 a month.

The facts, side by side

Each row shows the date it was checked and numbered links to the pages it was read from.

marks the side the facts favour on that row. Rows without one are a draw or a matter of fit.

Swipe the table sideways to see Gemini.

Claude and Gemini compared, fact by fact
ComparedClaudeAnthropicGeminiGoogle
Who makes it1913Anthropic. The chat app is Claude; the developer product is the Claude API.Google. The chat app is Gemini; the developer products are the Gemini API and Gemini Enterprise Agent Platform on Google Cloud.
Current models1914Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Claude Haiku 5.5Stable: Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite and 3.1 Flash-Lite. Preview: Gemini 3.1 Pro. Announced on 30 September 2026 and limited to the Fairwind Program: Gemini 4 Argon.
Where the vendor says to start19Opus 5.5 for most workloads; Fable 5.1 when Opus at higher effort still falls shortGemini 3.8 Flash or 3.5 Flash-Lite for any new project
Context window11011121M tokens on all four models1,048,576 input tokens on 3.8 Flash, 3.5 Flash-Lite and 3.1 Pro Preview
Longest single reply110111214Advantage: 128K tokens65,536 tokens on the same three. Gemini 4 Argon is announced with a 1M-token output limit.
Knowledge cutoff113Advantage: June 2026 on all fourMarch 2026 on Gemini 3.8 Flash. Google's model card adds that in some subjects its knowledge stops at January 2025.
What goes in and comes out1109Text and images in, text outAdvantage: Text, images, video, audio and PDF in, text out. Separate models generate images (Nano Banana 2.1), video (Gemini Omni Flash), music (Lyria 3.5), speech (Gemini 3.8 Flash TTS) and live voice (Gemini 3.8 Live).
Clouds you can buy it through113Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry, Claude Platform on AWSGoogle's own channels: the Gemini API and Gemini Enterprise Agent Platform on Google Cloud
Apps and coding agent382224252627Claude on web, iOS, Android and desktop. Claude Code in the terminal, IDE, desktop app and browser.Gemini on web, mobile, macOS and Windows. Gemini CLI in the terminal (open source, Apache 2.0), Google Antigravity as a desktop app, CLI and IDE, and Jules for coding tasks that run in the cloud.
API rate limits41617Published per tier. Start tier: 1,000 requests, 2,000,000 input tokens and 400,000 output tokens a minute on Opus 5.5, Sonnet 5.5 and Haiku 5.5, with a $500 monthly spend cap.Per-model limits are not published in the docs; each project sees its own in Google AI Studio. Published: Tier 1 has a $250 monthly spend cap and a $10 spend limit per 10 minutes.
Open-weight models121None listedAdvantage: Gemma 4 in five sizes (E2B, E4B, 12B, 26B A4B, 31B) under the Apache 2.0 licence
Training on your data73202322API and Team: not used by default. Free, Pro and Max list model training as opt-out.Paid API: not used. Unpaid API quota and Google AI Studio without a billing account: used, and human reviewers may read it. Gemini app: used while Keep Activity is on. Gemini Enterprise: not used.

Where each one is stronger

Claude Anthropic

  • Higher scores on the models you can buy

    Artificial Analysis scores Claude Opus 5.5 at 58 on its Intelligence Index (version 4.3.2, ten evaluations) at maximum effort, a run it labels "with fallback", and Gemini 3.8 Flash at 41 at high effort. Arena's agent leaderboard of 8 October 2026 lists Opus 5.5 first and Gemini 3.8 Flash 23rd. 282930

  • The cheaper small model

    Haiku 5.5 costs $0.10 input and $0.50 output per million tokens on prompts up to 100,000 tokens, against $0.30 and $2.50 on Gemini 3.5 Flash-Lite. Artificial Analysis scored Haiku 5.5 at 38 at high effort for $0.08 per task, and 3.5 Flash-Lite at 22 for $0.19. 21528

  • A longer runway

    Anthropic gives at least 60 days' notice before retiring a model and lists no retirement before 1 September 2027 for a current one. Google deprecated Gemini 3.7 Flash on 8 October 2026, eight weeks after releasing it, and now sends its requests to 3.8 Flash. 611819

Gemini Google

  • Video and audio in, media out

    Gemini 3.8 Flash accepts video, audio and PDF as well as text and images. The same API sells Nano Banana 2.1 for images at about $0.034 for a 1K image, Gemini Omni Flash for video at about $0.10 a second and Lyria 3.5 for music at $0.08 a song. 1015

  • A free tier, and a lower list price in the middle

    The Gemini API charges nothing on its free tier for 3.8 Flash and 3.5 Flash-Lite, in exchange for Google using the content. On the paid tier 3.8 Flash is $0.75 input and $3.75 output per million tokens until 31 December 2026, against $2 and $10 on Claude Sonnet 5.5. 15202

  • Open weights you can host

    Gemma 4 comes in five sizes from E2B to 31B under the Apache 2.0 licence. Anthropic lists no open-weight model. 211

Which to pick, by job

  • A coding agent that works through a large repository for hours

    Pick Claude

    Anthropic describes Opus 5.5 as built for long-running agentic coding. It scores 58 on the Artificial Analysis index and ranks first on the Arena agent leaderboard, where Gemini 3.8 Flash is 23rd. 12830

  • Pulling facts out of hours of video or recorded calls

    Pick Gemini

    Gemini 3.8 Flash takes video and audio directly, inside a 1,048,576-token input limit. Anthropic lists text and image input for Claude, so you would transcribe first with another service. 101

  • A prototype or side project with no budget

    Pick Gemini

    The Gemini API free tier covers 3.8 Flash and 3.5 Flash-Lite, and Gemini CLI allows 60 requests a minute and 1,000 a day on a personal Google account. Keep private data out: Google uses free-tier content to improve its products. 152520

  • A feature that classifies or extracts from short text thousands of times a day

    Pick Claude

    Haiku 5.5 costs $0.10 input and $0.50 output per million tokens; Gemini 3.5 Flash-Lite costs $0.30 and $2.50. Past 100,000 tokens a prompt, Haiku rises to $0.50 and $2.50 and the gap closes. 215

  • One paid chat app for writing, research and everyday questions

    Pick Gemini

    Google AI Plus is $4.99 a month with Gemini 3.1 Pro and 400 GB of storage; Claude Pro is $20. Choose Claude Pro instead if you want to write with Opus or work in Claude Code. 223

What it costs

List prices in US dollars from each vendor's own pages. API prices are per million tokens on the standard paid tier; both vendors take 50% off for batch jobs. A per-token price is not a per-task price: Google says Gemini 3.8 Flash can use more tokens on long tasks by design. Measure on your own prompts.

marks the side the facts favour on that row. Rows without one are a draw or a matter of fit.

Swipe the table sideways to see Gemini.

Claude and Gemini prices
ComparedClaudeAnthropicGeminiGoogle
Free plan322Claude Free: $0, with the Sonnet and Haiku modelsGemini Free: $0, with Gemini 3.5 Flash-Lite
Lowest paid plan322Claude Pro: $20 a month, or $17 a month billed annually ($200 up front)Advantage: Google AI Plus: $4.99 a month, with 3.6 Flash, 3.1 Pro and 400 GB of storage. Google AI Pro: $19.99 a month, with 3.8 Flash and 5 TB.
Heavy-use plan322Claude Max: from $100 a month, with 5x or 20x the usage of ProGoogle AI Ultra: $99.99 a month for 5x the usage of AI Pro, or $199.99 for 20x
Team seat322Team: $25 per seat a month, or $20 billed annually. Premium seat: $125, or $100 billed annually.Gemini Enterprise Business: from $21 per seat a month, up to 300 seats. Standard and Plus: from $30 per seat.
Top API model on sale21514Fable 5.1: $10 input, $50 output. Opus 5.5: $4 input, $20 output.Gemini 3.1 Pro Preview: $2 input, $12 output. Gemini 4 Argon is announced at $2 and $10 but is not on general sale.
Mid API model215Sonnet 5.5: $2 input, $10 outputAdvantage: Gemini 3.8 Flash: $0.75 input, $3.75 output until 31 December 2026; $1.50 and $7.50 from 1 January 2027
Cheapest API model215Haiku 5.5: $0.10 input, $0.50 output for prompts up to 100,000 tokens; $0.50 and $2.50 above thatGemini 3.1 Flash-Lite: $0.25 input, $1.50 output. Gemini 3.5 Flash-Lite: $0.30 input, $2.50 output.
Free API tier21520None. New users get a small amount of free credit.Advantage: Yes, on 3.8 Flash, 3.6 Flash and 3.5 Flash-Lite. Google uses what you send on it to improve its products. Gemini 3.1 Pro Preview is paid only.
Long prompts215Standard rate across the full 1M window on Fable 5.1, Opus 5.5 and Sonnet 5.5One rate at any length on 3.8 Flash and the Flash-Lite models. Gemini 3.1 Pro Preview charges $4 input and $18 output once a prompt passes 200,000 tokens.
Cached input215Fable 5.1: $0.25. Opus 5.5: $0.20. Sonnet 5.5: $0.10. Haiku 5.5: $0.01.Gemini 3.1 Pro Preview: $0.20. Gemini 3.8 Flash: $0.075. Gemini 3.5 Flash-Lite: $0.03. Stored caches also bill by the hour: $0.50 per million tokens on 3.8 Flash. Both 3.8 Flash prices double on 1 January 2027.

What changed recently

  1. Scheduled: Gemini 3.8 Flash and 3.6 Flash rise from $0.75 and $3.75 per million tokens to $1.50 and $7.50. 15

  2. Scheduled: Anthropic retires Claude Sonnet 4.5 on the Claude API; Sonnet 5.5 replaces it. 6

  3. Google deprecated Gemini 3.7 Flash and 3.5 Flash and now routes their requests to 3.8 Flash and 3.6 Flash. 18

  4. Anthropic launched Claude Haiku 5.5 and cut Sonnet 5.5 cache reads from $0.20 to $0.10. 5

  5. Google announced Gemini 4 Argon at an introductory $2 and $10 per million tokens, for Fairwind Program members first. 14

  6. Anthropic launched Claude Opus 5.5 at $4 and $20, down from $5 and $25 on Opus 5. 5

  7. Google released Gemini 3.8 Flash. 18

  8. Anthropic launched Claude Fable 5.1 at $10 and $50, with cache reads at $0.25. 5

Questions people ask

Is Claude better than Gemini?

For writing, coding and agent work, Claude ranks higher today. At high effort Artificial Analysis scores Claude Opus 5.5 at 54 and Sonnet 5.5 at 47, against 41 for Gemini 3.8 Flash; Gemini 3.1 Pro Preview scores 30. Gemini is the better choice for video and audio input, generated media and a free tier. 281015

Which is cheaper, Claude or Gemini?

Gemini is cheaper in the middle and Claude at the bottom. Gemini 3.8 Flash is $0.75 input and $3.75 output per million tokens until 31 December 2026, then $1.50 and $7.50; Claude Sonnet 5.5 is $2 and $10. Claude Haiku 5.5 is $0.10 and $0.50 on prompts up to 100,000 tokens; Gemini 3.5 Flash-Lite is $0.30 and $2.50. List price is not task cost: Artificial Analysis measured $1.24 per task for 3.8 Flash and $0.88 for Sonnet 5.5, both at high effort. 15228

What is Gemini 4 Argon, and can I use it?

Not yet, unless you are in the Fairwind Program. Google announced Gemini 4 Argon on 30 September 2026 and is releasing it to trusted cyber defenders first, then to paid API customers and Google AI Ultra subscribers, with no date given. The announced price is $2 input and $10 output per million tokens, rising to $4 and $20 after an introductory period. Artificial Analysis lists it at 53, behind Claude Opus 5.5 at 58. 1428

What is the context window of Claude and Gemini?

About a million tokens on both. Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 each take 1M tokens. Gemini 3.8 Flash, 3.5 Flash-Lite and 3.1 Pro Preview each accept 1,048,576 input tokens. The difference is the reply: Claude can write 128K tokens in one response and these Gemini models 65,536. 1101112

Do Claude and Gemini train on what I type?

Not on paid API usage: Anthropic does not train on inputs or outputs from its commercial products by default, and Google does not use paid Gemini API content to improve its products. Google does use content sent on the free Gemini API tier, and human reviewers may read it. In the apps it is a setting: Claude lists model training as opt-out on Free, Pro and Max, and the Gemini app trains on chats while Keep Activity is on. 720323

Should I pay for Claude Pro or Google AI Pro?

They cost the same: Claude Pro is $20 a month and Google AI Pro is $19.99. Claude Pro adds the Opus models, Claude Code and Research. Google AI Pro adds Gemini 3.8 Flash, higher limits in Jules and Google Antigravity, and 5 TB of storage. Pick Claude Pro to write and code with Claude, and Google AI Pro if you work in Gmail, Docs and Drive. 322

Sources

Each page was read on the date beside it. The links open the current version, which may have changed since.

  1. 1Anthropic: Claude models overview
  2. 2Anthropic: Claude API pricing
  3. 3Anthropic: Claude plans and pricing
  4. 4Anthropic: Claude API rate limits
  5. 5Anthropic: Claude Platform release notes
  6. 6Anthropic: model deprecations
  7. 7Anthropic: is my data used for model training (commercial products)
  8. 8Anthropic: Claude Code overview
  9. 9Google: Gemini API models
  10. 10Google: Gemini 3.8 Flash model page
  11. 11Google: Gemini 3.5 Flash-Lite model page
  12. 12Google: Gemini 3.1 Pro Preview model page
  13. 13Google DeepMind: Gemini 3.8 Flash model card
  14. 14Google: Gemini 4 Argon announcement
  15. 15Google: Gemini API pricing
  16. 16Google: Gemini API rate limits
  17. 17Google: Gemini API billing
  18. 18Google: Gemini API release notes
  19. 19Google: Gemini API deprecations
  20. 20Google: Gemini API additional terms of service
  21. 21Google: Gemma 4 model card
  22. 22Google: Gemini subscriptions (United States)
  23. 23Google: Gemini Apps Privacy Hub
  24. 24Google: Gemini app release notes
  25. 25Google: Gemini CLI readme
  26. 26Google: Antigravity pricing
  27. 27Google: Jules
  28. 28Artificial Analysis: LLM leaderboard
  29. 29Artificial Analysis: Intelligence Index methodology
  30. 30Arena: agent leaderboard

Take the free Starter Kit.

Five working files from the IImagined programs: 32 Instagram hooks, an AI character template, a pricing worksheet and two calculators. Free, and they open in your browser.

Get the free Starter Kit

Related program

AI SaaS Builder. Validate an idea, build on Supabase and Next.js, add AI with the Claude API, use Claude Code and MCP, deploy on Vercel, launch and bill with Stripe. Included in All Access.

IImagined publishes this comparison and sells the All Access membership.