Technology

LLM API Market 2024: Google Leads as OpenAI Slips

## LLM API Market 2024: Key Trends and Model Leaderboard As we reach the midpoint of 2024, the competitive landscape for large language models (LLMs) is shifting rapidly. The so-called "LLM Wars" ar...
Noll
4 min read
#LLM API market#large language models#OpenRouter#token usage

LLM API Market 2024: Key Trends and Model Leaderboard

As we reach the midpoint of 2024, the competitive landscape for large language models (LLMs) is shifting rapidly. The so-called "LLM Wars" are intensifying, with major changes in the LLM API market for text-generation technologies. Is OpenAI losing its lead? Which large language models are developers actually using, and what trends are shaping the global LLM API market?

A recent analysis by X user "karminski"—using comprehensive data from the API aggregator OpenRouter—provides valuable insights into token usage, market share, and developer preferences for the first half of 2024.

Source: https://x.com/karminski3/status/1942612077241311386

Let's break down the findings from the OpenRouter dataset.

What is OpenRouter?

OpenRouter is a widely used API aggregator that allows developers to access and compare various large language models through a unified interface. Its usage data offers a unique perspective on real-world developer preferences and LLM API market trends.

Model Leaderboard: Most Popular LLMs Among Developers

Which large language models are leading the API market in 2024? According to the latest OpenRouter data:

  • Google's Gemini-2.0-Flash is the most used model, topping the leaderboard.
  • Anthropic's Claude-Sonnet-4 holds second place.
  • Google's Gemini-2.5-Flash-Preview-0520 rounds out the top three.
  • DeepSeek V3 (both free and paid versions) ranks 4th and 5th; combined, DeepSeek V3 rivals the second-place model in total API usage.

Trend chart of the most popular models

LLM API Market Share Breakdown

The LLM API market share data reveals:

  • Google leads with a commanding 43.1% market share.
  • DeepSeek and Anthropic follow closely, with 19.6% and 18.4% respectively.

LLM market share breakdown

Developer Preferences by Domain

OpenRouter's data also highlights which LLMs developers prefer for specific use cases.

Usage data for domain-specific models

Programming Domain: Top LLMs for Coding

For code generation and programming tasks, developers increasingly choose Anthropic's Claude series. These models are favored for their advanced capabilities in the programming domain.

Model calls in the programming domain

Translation Domain: LLMs for Language Translation

In translation, the LLM API market is more fragmented. Google's models stand out, leveraging extensive linguistic datasets to provide robust translation capabilities.

Model calls in the translation domain

Role-Playing Applications: Creative LLM Usage

For creative and role-playing applications, developers experiment with various LLMs to balance creativity, coherence, and cost for narrative-driven apps.

Model calls in the role-playing domain

Advertising Domain: LLMs for Marketing Content

In marketing, developers use a range of LLMs to generate ad copy and campaign content, optimizing for creativity and cost-effectiveness.

Model calls in the marketing domain

Developer Interfaces: How LLMs Are Accessed via OpenRouter

Examining how developers access LLMs through OpenRouter reveals integration patterns and workflow preferences.

OpenRouter interface usage trends

Key Takeaways: 2024 LLM API Market Insights

  • Google's Strategic Dominance: Google leads the LLM API market, capturing nearly half of all usage. Its Gemini-2.0-Flash model is especially popular due to its cost-effectiveness and performance.
  • Anthropic's Coding Focus: Anthropic's Claude-3.5, 3.7, and 4 models are gaining traction in the programming domain, appealing to developers seeking specialized coding capabilities.
  • OpenAI's Decline: OpenAI's lower API market share may result from strategic decisions, AccessKey requirements, or less competitive pricing for high-volume use.
  • DeepSeek's Rise: DeepSeek V3 is a 2024 success story, with strong developer retention. Its faster response times and lower time to first token (TTFT) make it more attractive than DeepSeek-R1.
  • Meta's Llama Fades: The Llama series is losing momentum in the LLM API market, with declining developer interest.
  • Emerging Players: Mistral AI holds 3% market share, especially among European developers interested in fine-tuning open-source models. X-AI's Grok and Alibaba's Qwen (Tongyi Qianwen) are present but face challenges in gaining broader adoption.

Glossary:

  • TTFT (Time to First Token): Measures how quickly an LLM starts generating output.
  • SOTA (State-of-the-Art): Refers to the leading models in a specific field.

For more on LLM performance, see How to Choose the Best LLM for Your Application and OpenRouter API Integration Guide.

Related Articles

Technology
6 min

SFT Flaw: A Learning Rate Tweak Unlocks LLM Potential

Discover a critical flaw in Supervised Fine-Tuning (SFT) that limits LLM performance. Learn how a simple learning rate tweak unifies SFT and DPO for a 25% gain.

Noll
Supervised Fine-Tuning (SFT)Direct Preference Optimization (DPO)+2 more
Technology
7 min

Two Major Challenges in Reinforcement Learning Finally Solved by ICLR Papers

Traditional reinforcement learning models struggle with real-time applications due to "AI lag." Two ICLR 2025 papers from Mila introduce groundbreaking solutions to tackle inaction and delay regret, enabling large AI models to operate in high-frequency, dynamic environments without compromising speed or intelligence.

Noll
TechnologyAI+1 more
Technology
13 min

Discuss the infrastructure requirements of Agentic AI.

The rise of Agentic AI places unprecedented demands on our infrastructure. This article explores the emerging software and hardware requirements, from specialized runtimes and memory services to zero-trust security models, dissecting AWS's new Bedrock AgentCore platform and discussing the future of AI infrastructure.

Noll
TechnologyAI+1 more

About This Article

Topic: Technology
Difficulty: Intermediate
Reading Time: 4 minutes
Last Updated: July 9, 2025

This article is part of our comprehensive guide to Large Language Models and AI technologies. Stay updated with the latest developments in the AI field.

All Articles
Share this article to spread LLM knowledge