Side-by-side comparison · AI Assistants & LLMs
DeepSeek vs Kimi vs Perplexity
Everything that matters when choosing between them, in one place: our rating, what each can do, what they cost, and the trade-offs.
Comparing 3 tools, the maximum.
At a glance
Pick it if: Anyone who wants a strong reasoning assistant for free, and developers watching their API costs.
Pick it if: People who want an agent to produce finished slides, documents, spreadsheets and websites, and developers after a cheaper coding model.
Pick it if: Anyone who researches as part of their job and is sick of having ten tabs open.
Perplexity scores highest overall, but DeepSeek gives you more for your money. Which one is right comes down to what you value most.
Overview
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| In one line | A free Chinese assistant that's strong at maths, logic and code. | Moonshot AI's assistant with a strong agent for slides, documents, sheets and websites. | Ask a question and get a short answer from the web, with links to every source. |
| Pricing model | Free | Freemium | Freemium |
| Starting price | Pay per use | $19/mo or $15/mo | $17/mo |
| Free plan | Yes | Yes | Yes |
| Works with Claude Code | No | No | Yes |
Rating breakdown
Our scores out of 5. The higher bar in each row is highlighted.
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Overall rating | 3.8 | 3.8 | 4.2Winner |
| Ease of useHow quickly someone non-technical gets value from it. | 3.5 | 4.0 | 4.5 |
| FeaturesHow completely it covers the job it's built for, and how well the pieces fit together. | 3.0 | 4.5 | 3.5 |
| Output qualityHow good the results are before you fix them. | 4.0 | 4.0 | 4.5 |
| Value for moneyWhat you get for the price, including the free plan. | 5.0 | 3.5 | 4.0 |
| Speed & reliabilityHow fast it is and how often it falls over. | 3.5 | 3.0 | 4.5 |
Feature comparison
A tick means it does this today. A dash means we haven't checked.
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Access & platform | |||
| Free plan | Yes | Yes | Yes |
| Mobile app | Yes | Yes | Yes |
| Desktop app | No | Yes | Yes |
| Browser extension | No | No | Yes |
| API access | Yes | Yes | Yes |
| Teams & business | |||
| Team workspace | No | – | Yes |
| Enterprise plan or SSO | No | No | Yes |
| Connects to other apps | No | No | Yes |
| Open source or self-hosted | Yes | Yes | No |
| AI abilities | |||
| Works with your files | Yes | Yes | Yes |
| Understands images | Yes | Yes | Yes |
| Makes images | No | – | Yes |
| Makes video | No | No | Yes |
| Writes code | Yes | Yes | Yes |
| Runs agents or automations | No | Yes | Yes |
| Learns your brand or style | No | No | No |
| Works in many languages | Yes | Yes | Yes |
| Capabilities score | 8/ 17 | 10/ 17 | 15/ 17 |
Pricing
Plans as listed on each vendor's site when we last checked. Confirm before you buy.
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Plans | Chat appOur pick Free Anyone who wants a capable reasoning assistant at no cost. | Adagio Free Trying the chat and a taste of the agent. | Free Free Everyday questions and quick fact-checks. |
API (deepseek-flash) Pay per use, from $0.15 per million input tokens Developers building cheap features into their own apps. | PlusOur pick $19/mo or $15/mo billed yearly Anyone who wants to use the agent most weeks. | ProOur pick $17/mo billed yearly Anyone who researches as part of their job. | |
API (deepseek-v4-pro) Pay per use, from $0.66 per million input tokens Developers who want the stronger model. | Pro $39/mo or $31/mo billed yearly Regular users who run out of credits on Plus. | Max $167/mo billed yearly Power users running the Computer agent all day. | |
Max $99/mo or $79/mo billed yearly Heavy users running several jobs at once. | Enterprise Pro $34/seat/mo billed yearly Teams that need admin controls and data protection. | ||
Ultra $199/mo or $159/mo billed yearly Power users running agents all day. | Enterprise Max $271/seat/mo billed yearly Research teams working with large datasets and files. | ||
| Which plan to pick | The app is free, so the only decision is whether the privacy trade-off is acceptable for what you're asking. If you need the model inside your own software, have a developer use the API or run the open model through a provider that keeps your data in Australia. | Try the free plan to see the agent's output. If you like it, Plus at $19 is the sensible plan, and the yearly price drops it to $15. Only move up if you keep running out of credits, which long Swarm jobs do quickly. | Start free and see how often you hit the Pro search limit. If you research most days, Pro pays for itself in saved tabs. Its 4,000 bonus Computer credits are a one-off and run out after about 30 days, so don't count them as a monthly allowance. Max only makes sense if you're using the Computer agent heavily enough to burn through the credits. |
Key features
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Standout features |
|
|
|
Pros and cons
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Pros |
|
|
|
| Cons |
|
|
|
Who each one is for
| Attribute | DeepSeek | Kimi | Perplexity |
|---|---|---|---|
| Best for |
|
|
|
| You'll find it used by |
|
|
|
Questions people ask
Which is better, DeepSeek, Kimi or Perplexity?
Perplexity scores highest overall in our ratings (DeepSeek 3.8/5, Kimi 3.8/5 and Perplexity 4.2/5). DeepSeek is best for Maths, logic and step-by-step problems and coding help. Kimi is best for turning a brief into a slide deck and research reports and document drafts. Perplexity is best for fact-checking and quick research and competitor and market research.
Which is cheaper, DeepSeek, Kimi or Perplexity?
DeepSeek: There's a free plan, and paid plans start at Pay per use. Kimi: There's a free plan, and paid plans start at $19/mo or $15/mo. Perplexity: There's a free plan, and paid plans start at $17/mo.
Which is easier to use, DeepSeek, Kimi or Perplexity?
Perplexity rates highest for ease of use in our scores. Anyone who researches as part of their job and is sick of having ten tabs open.
How we compare
Ratings are our own, from using each tool, across ease of use, features, output quality, value for money, speed & reliability.
Swap in another tool
More tools from AI Assistants & LLMs in the library.
