Eduardo Ordax - The 2025 Landscape of LLMs — Updated View of the Big Players in the Game of AI - May 2025

Free

The 2025 Landscape of LLMs — Updated View of the Big Players in the Game of AI

FreeFree tier

Type

Open Source

Visit Website GitHub (38 stars)

About Eduardo Ordax - The 2025 Landscape of LLMs — Updated View of the Big Players in the Game of AI - May 2025

This LinkedIn post by Eduardo Ordax presents the 2025 landscape of large language models (LLMs), offering an updated view of major AI research labs, their latest models, and how those models can be accessed. It covers approximately 95% of real-world LLM usage and highlights key trends: no single clear front-runner, growing customer demand for model choice and interoperability, the rise of reasoning-first models fueling agentic AI, proprietary models still leading but open-source closing the gap, U.S. leadership with increasing international competition, dominance of cloud and API access, serverless as the default deployment model, and the market consolidation with everyone else comprising less than 5%.

Key Features

Covers major AI research labs and their latest models

Focuses on model accessibility via APIs and cloud platforms

Highlights convergence of model performance and importance of orchestration

Discusses rise of reasoning-first models and agentic AI architectures

Compares proprietary vs open-source adoption and market share

Provides observations on geographic leadership (US vs international)

Notes serverless as default deployment preference

States the post captures about 95% of real-world LLM usage

Pros & Cons

Pros

Provides a concise, high-level overview of the 2025 LLM market
Updates a previous version from 18 months ago, showing evolution
Cites real-world usage coverage of ~95%
Identifies actionable trends like model choice and reasoning-first models
Offers accessible insights for both technical and non-technical audiences

Cons

Not an exhaustive list of every LLM — deliberately selective
Based on personal opinion and observation rather than quantitative analysis
Limited depth on individual models or technical benchmarks
Only a single LinkedIn post, not a formal research report

Best For

Understanding the competitive landscape of LLMs in 2025Guiding model selection decisions for AI practitioners and enterprisesAnalyzing market trends for investment or strategy planningEvaluating model access and interoperability optionsTracking the shift from model performance to ecosystem and routing

FAQ

What models are covered in this landscape?

The post focuses on leading AI research labs and their latest models, covering about 95% of what is being used in real-world scenarios today. It does not list every single LLM.

How does this 2025 landscape differ from the version 18 months ago?

The author notes that there is no longer a clear front-runner; the market has become a fairly even playing field where model differences are small for most use cases. Customer demand for model choice, interoperability, and evaluation frameworks has grown.

What are the main takeaways from the landscape?

Key takeaways include: no clear leader, model choice is the new normal, reasoning-first models are rising, proprietary still leads but open-source is catching up, US ahead but international competition is heating up, cloud/API access dominates, and serverless is the default deployment.

Is this landscape based on quantitative data?

No, it is based on the author's observations and experience in the AI space. It is a personal analysis of the market trends.