DeepSeek Undercuts Rivals With $0.75 Search API

Overview

DeepSeek has launched its new Search and Fetch APIs at a price point that dramatically undercuts competitors in the artificial intelligence space. The Chinese AI firm announced that its Search API and Fetch API are now available to the public, priced at just $0.75 per 1,000 calls.

This pricing strategy represents a significant disruption in the market for retrieval-augmented generation (RAG) infrastructure. For context, major rivals charge substantially higher rates: OpenAI and Anthropic both list their equivalents at $10 per 1,000 calls, Grok charges $5 per 1,000 calls, and Google’s Gemini services are priced at $14 per 1,000 calls.

The move positions DeepSeek as a cost-efficient alternative for developers building agent-based applications and systems requiring real-time web data. By leveraging its own in-house search engine, the company aims to provide a scalable backend for the growing demand for live internet access in AI workflows.

Pricing and Market Impact

The financial disparity between DeepSeek’s offering and established players is stark. At $0.75 per 1,000 calls, DeepSeek is approximately 13 times cheaper than OpenAI and Anthropic, and nearly 19 times cheaper than Google’s Gemini API. For high-volume applications, such as enterprise-grade chatbots or automated research tools, this difference can translate into massive cost savings.

• DeepSeek: $0.75 per 1,000 calls
• Grok: $5 per 1,000 calls
• OpenAI & Anthropic: $10 per 1,000 calls
• Google Gemini: $14 per 1,000 calls

"DeepSeek Undercuts Rivals With $0.75 Search API"

Industry analysts suggest that this aggressive pricing could force a reevaluation of costs across the sector. As more companies integrate real-time search capabilities into their large language model (LLM) applications, the operational expenditure associated with fetching live data becomes a critical factor. DeepSeek’s low margin per request suggests a volume-driven strategy, aiming to capture market share by making advanced RAG tools accessible to smaller teams and individual developers who previously found competitor pricing prohibitive.

The announcement was made via the official DeepSeek X account, emphasizing that the infrastructure is designed to handle hundreds of millions of requests daily worldwide. This scalability claim is crucial for potential enterprise adopters concerned about reliability and uptime.

Features and Integration

The new APIs provide agents and RAG applications with direct access to live web content through two distinct endpoints: the Search API and the Fetch API.

The Search API allows developers to query DeepSeek’s proprietary search engine, returning relevant results based on real-time data. The Fetch API complements this by enabling the extraction and processing of content from specific URLs, which is essential for grounding AI responses in current facts rather than static training data.

These tools are particularly valuable for:

1. AI Agents: Systems that need to perform multi-step reasoning involving live data retrieval.
2. RAG Applications: Tools that require up-to-date information to supplement the knowledge contained within an LLM’s parameters.
3. Research Automations: Workflows that aggregate and synthesize information from across the web.

Integration is streamlined for users of popular developer environments. According to the announcement, Search and Fetch calls made via server-side tools in DeepSeek Harness, Claude Code, and Codex are currently free for DeepSeek API users. This incentivizes adoption within existing developer ecosystems while the broader public rollout begins.

Documentation for the new features is available at the official DeepSeek API guides, providing technical specifications for implementation.

Timeline and Availability

While the APIs are now open to everyone, there is a specific timeline regarding billing for certain integrated tools. The free usage period for calls made through DeepSeek Harness, Claude Code, and Codex will end on October 31, 2026, at 16:00 UTC. After this date, standard billing rates will apply to these specific integration paths.

For general API usage, the $0.75 per 1,000 calls rate applies immediately to all users. This structure allows developers to test and integrate the tools during the current free window before costs are incurred, reducing the barrier to entry for experimentation.

The timing of this launch coincides with an intensifying competition among AI providers to offer comprehensive "full-stack" solutions. Having a built-in, cost-effective search layer is becoming a differentiator, as it removes the need for developers to stitch together separate search providers like Google Custom Search or SerpAPI, which often come with higher latency and price tags.

DeepSeek’s ability to offer this at such a low cost is attributed to its investment in its own search infrastructure. By owning the underlying technology rather than relying on third-party APIs, the company can maintain tighter control over costs and performance, passing those savings on to its customers.