- Anthropic now leads the enterprise LLM market with 32% usage, ahead of OpenAI at 25%.
- Claude dominates in AI coding tools, capturing 42% of developer preference.
- Performance, not cost, is driving businesses to adopt newer, more powerful models.
- Open-source LLMs are fading in enterprise use, dropping from 19% to 13% in six months.
Anthropic has emerged as the top large language model (LLM) provider for businesses, edging past industry heavyweight OpenAI. According to a new report by Menlo Ventures, Anthropic now holds a commanding 32% share of enterprise LLM usage, leaving OpenAI at 25%, Google at 20%, and Meta’s Llama trailing at 9%.
Despite Menlo Ventures’ significant investment in Anthropic, the findings echo broader industry sentiment: when it comes to powering real-world applications, Claude, Anthropic’s family of AI models, is leading the pack.
Claude’s Secret Weapon? Code Generation Dominance
A major reason behind Anthropic’s rise is its success in the world of code generation. The report identifies programming as the most impactful enterprise use case for AI to date, and Claude is dominating that niche. With 42% of developers preferring Claude for coding tasks, compared to just 21% using OpenAI, Anthropic has firmly planted its flag in the developer community.
Over the past year, Claude has powered everything from AI-enhanced development environments, such as Cursor and Windsurf, to enterprise-grade agents, including Claude Code and All Hands. This ecosystem has helped spawn an estimated $1.9 billion in economic activity, just from GitHub Copilot’s Claude integration alone.
Training That Actually Works in the Real World
What’s helping Claude win over engineers and enterprises alike is its unique approach to training. Anthropic uses something called reinforcement learning with verifiable rewards (RLVR).
Unlike traditional reward models based on subjective feedback, RLVR uses binary scoring, 1 if the AI-generated code works, 0 if it doesn’t. It’s an approach tailor-made for the precision demands of programming.
Claude’s ability to reason through complex tasks step-by-step, combined with its seamless use of external tools via the open Model Context Protocol (MCP), makes it ideal for real-world business use.
By accessing calculators, code libraries, and web search tools in real-time, Claude delivers faster, more accurate results across a wide range of tasks.
Performance Beats Price in the AI Race
Perhaps the most interesting insight from the Menlo Ventures study is that cost isn’t the main concern for businesses choosing AI models, performance is.
Even when LLM pricing drops by a factor of 10, companies don’t opt for older, cheaper models. Instead, they migrate en masse to the best-performing solution.
This trend is contributing to the rapid churn in AI model adoption. As new releases continue to deliver big gains in accuracy, speed, and reliability, businesses are willing to pay for better performance, especially when it leads to faster development cycles and fewer errors in production.
Open-Source Models Losing Ground
While new open-source models continue to enter the market, they’re struggling to gain traction. Six months ago, open-source LLMs accounted for 19% of enterprise workloads. Today, that number has dropped to just 13%.
Even Meta’s Llama, arguably the most prominent “open” model, hasn’t changed the tide. Many businesses remain skeptical about open-source LLMs, citing lower performance, limited support, and concerns around models developed by Chinese companies.
The promise of flexibility and on-premise deployment hasn’t been enough to outweigh concerns about reliability.
Where Do We Go From Here?
The enterprise AI market is evolving fast. With 74% of startups and 49% of large enterprises now running AI in production, not just testing or building models—the pressure is on to choose the best-performing tools.
“Predicting where AI is headed next is tricky,” Menlo Ventures notes. “But it’s clear the building blocks for the next wave of transformative AI businesses are being laid today.”
For now, Anthropic has taken the lead. Whether it can hold onto that advantage in a constantly shifting industry remains to be seen.
Stay tuned. The AI era is only just beginning.
Follow TechBSB For More Updates
