How Makers Build
Check my stackPlan my stack
Menu

Claude vs DeepSeek

Two model provider options for LLM API. When each fits, what it costs, who moves from one to the other, and what makers who chose it say.

ClaudeModel providerllms.txtIn Lovable, Replit
DeepSeekModel provider
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

Which fits you

Choose Claude if
  • You want the strongest general models and the widest ecosystem of examples and integrations

Use it whenYour feature is an agent that calls tools, writes code or works through long documents.

Trade-offNo image generation or embedding models of its own, so those need a second provider.

Choose DeepSeek if
  • Token cost dominates your budget, for example high-volume batch or agent workloads

Use it whenToken cost dominates your budget, for example high-volume batch or agent workloads.

Trade-offServed from China, which matters for some customers' data requirements; peak-hour prices are double off-peak.

At a glance

ClaudeDeepSeek
Used by251 makers' products · 285 open-source projects11 makers' products · 24 open-source projects
Cost at default usageinput tokens 50 million tokens, output tokens 10 million tokens$10/mo Claude Haiku 5.5$27/mo deepseek-flash
Moved to it on GitHubpull requests since Oct 20248 from DeepSeek17 from Claude
Downloads31.6M/wk4.1× vs npm1.9M/wk4.9× vs npm
PricingPay per token. · paid from Pay per tokenPay per token from a prepaid balance; off-peak hours are half price. · paid from Pay per token
Free tierNoNo
Open sourceNoNo
Incidents, 90 daysfrom its status page50+ (21 major)no public status feed

Cost as you grow

At 1M tokens Claude costs less ($0.20 vs $0.54); and still does at 5B tokens ($1,000 vs $2,700).

$0$100$500$1,000$2,0001510501005001k5k
DeepSeekClaudex: input tokens per month (million tokens), other usage scaled with it · cheapest usable plan at each point, list prices · try your own numbers
The numbers, plan by plan
Input tokens per monthClaudeDeepSeek
1$0.20 Claude Haiku 5.5$0.54 deepseek-flash
5$1.00 Claude Haiku 5.5$2.70 deepseek-flash
10$2.00 Claude Haiku 5.5$5.40 deepseek-flash
50$10 Claude Haiku 5.5$27 deepseek-flash
100$20 Claude Haiku 5.5$54 deepseek-flash
500$100 Claude Haiku 5.5$270 deepseek-flash
1,000$200 Claude Haiku 5.5$540 deepseek-flash
5,000$1,000 Claude Haiku 5.5$2,700 deepseek-flash

From each vendor's pricing page: Claude, DeepSeek.

Who moves from one to the other

Public pull requests on GitHub since Oct 2024 whose title says "Claude to DeepSeek" or the reverse — real code changes, by developers in general rather than makers only.

DeepSeek → Claude8 PRs
All matching pull requests on GitHub ↗
Claude → DeepSeek17 PRs
All matching pull requests on GitHub ↗

What makers say

Makers on using it for LLM API, from Product Hunt and Starter Story interviews, each linked to the source. Products with a page of their own and fuller notes first.

On Claude
Taskade Genesis works beautifully with Claude. It brings strong reasoning and long-range thinking to any Genesis app, helping systems plan, adapt, and make sense of complex inputs.
Taskade, the makerSep 2026 ↗
We're excited about Claude-3.5 Sonnet and the upcoming 3.5 Opus. We'll introduce Opus to AICamp users upon release. We use Sonnet for its superior pair programming capabilities.
AICamp, the makerSep 2026 ↗
Huge shoutout to the Anthropic team — their Claude models are the backbone of some of our agents. Fast, reliable, and beautifully open‑source. Exactly what the AI world needs more of.
Calk AI, the makerSep 2026 ↗
149 more on the Claude page →
On DeepSeek
DeepSeek delivers serious reasoning at a fraction of the cost of comparable models. We chose it because powerful AI shouldn't have a premium price tag attached to every use case.
IFTTT, the makerSep 2026 ↗
DeepSeek Chat serves as our reliable foundation for basic content creation tasks. Its efficient performance and consistent output quality make it perfect for handling straightforward development requests.
MGX (Now Atoms), the makerSep 2026 ↗
DeepSeek gives us frontier-level reasoning at a fraction of the cost, which means we can run AI features at scale without watching the meter. Perfect for the high-volume inference paths where every token counts.
Weavable, the makerSep 2026 ↗
4 more on the DeepSeek page →

Loved and watch-outs

Themes that recur in makers' words and Hacker News comments, each linked to what it summarises.

Claude
Most loved
  • It is strong at writing code and powers many code generation and coding agent products. PHPH 2PH 3PH 4
  • It stays coherent through long, multi-step reasoning and agent workflows. PHPH 2PH 3PH 4
  • Its writing sounds natural and matches a user's tone rather than reading as AI-generated. PHPH 2PH 3PH 4
Watch-outs
  • Outages are frequent, with uptime below three nines across the API and apps. HNHN 2HN 3HN 4
  • API pricing is a large multiple of subscription pricing, making heavy agent use expensive. HNHN 2HN 3
  • Postmortems confirmed periods of degraded output, including a lowered default reasoning effort in Claude Code. HN
On Product Hunt: 5.0★, 1,012 reviews · mentioned most: AI assistant, code generation, contextual understanding · complaints: message limits, typography issues
DeepSeek
Most loved
  • It delivers strong reasoning at a small fraction of frontier model prices. PHPH 2HNHN 2
  • Recent Flash models handle real coding work in agent harnesses close to frontier models. HNHN 2HN 3HN 4
  • It has fewer guardrails, so it helps with security and malware analysis other models refuse. HNHN 2HN 3HN 4
Watch-outs
  • It burns many thinking tokens, so cost per finished task can exceed cheap per-token prices. HNHN 2HN 3HN 4
  • It hallucinates, ignores instructions and struggles with end-to-end builds that frontier models finish. HNHN 2HN 3HN 4
On Product Hunt: 4.9★, 53 reviews · mentioned most: deep thinking ability, supports coding tasks, handles complex tasks · complaints: slow performance, not as good as ChatGPT

Who uses each

Claude251 makers' products
8baseAccordio AIAgentcardAgentplace.ioAgentXAI Code Reviews by MatterAIAI EmailyAICampAirial TravelAirKarenAnythingAstrid+239 more
DeepSeek11 makers' products
Genstore.aiIFTTTMGX (Now Atoms)Read & Give by BonoShipable AI by CNTXT AIWeavableWebhoundBlink.newChat DataBrowser UseCaptioner
Used by both — often one replacing the other, or each for a different part of the product
Blink.newGenstore.aiMGX (Now Atoms)Shipable AI by CNTXT AIWeavableChat Data

What makers pair each with

With Claude
Transactional Email
Frontend Framework
Hosting
Payments
With DeepSeek
Frontend Framework
Transactional Email
Hosting
Database

Neither quite right?

See the whole LLM API decision →
OpenAI APIModel providerThe broadest model lineup — text, images, speech, embeddings — and the largest ecosystem.vs Claude →vs DeepSeek →
Gemini APIModel providerVery long context and multimodal input, with a free tier to start.vs Claude →vs DeepSeek →
GroqOpen models, hostedVery low-latency inference on open models, for real-time features.vs Claude →
Mistral AIModel providerA European provider whose API covers chat, coding, OCR and speech models.vs Claude →vs DeepSeek →
OpenRouterGateway (many providers, one API)Trying many models from many providers with one API key and one bill, without opening an account at each.vs Claude →
LiteLLMGateway (many providers, one API)Running your own OpenAI-compatible proxy in front of your own provider keys, with retries, routing and per-key budgets.