The enterprise AI market has its first genuine platform fork. Anthropic and OpenAI are competing for the same buyers, financial services teams, legal departments, enterprise operations, with products built on meaningfully different bets about how AI value gets delivered at scale.
Anthropic launched Claude Opus 4.6 on February 5, 2026, positioning it explicitly for knowledge work: financial analysis, regulatory document review, research, and productivity tasks involving spreadsheets and presentations. The model carries a 1M token context window in beta, a meaningful differentiator for enterprises processing large document sets, and is available across Claude’s own platform, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry. API pricing is $5 per million input tokens and $25 per million output tokens.
Around the same time, OpenAI launched Frontier, an enterprise platform built not around a single model but around agent deployment infrastructure: shared context management, onboarding workflows, and permission frameworks for businesses running AI agents across their systems. Frontier’s pricing hasn’t been publicly disclosed. That asymmetry, one vendor competing on model performance and transparent pricing, the other on platform completeness, defines the current competitive split.
One concrete signal of Anthropic’s traction: Claude Code, its developer-facing coding product, reached $1 billion in annualized run-rate revenue within approximately six months of its public launch, per Anthropic. Run-rate revenue reflects the current pace of bookings projected forward over twelve months, it is not cumulative recognized revenue, but the figure is notable for a product that didn’t exist a year ago. Multiple industry observers note that Anthropic and OpenAI are increasingly competing for the same enterprise buyers, though the competitive framing comes primarily from analyst and trade coverage rather than either company’s official positioning.
Anthropic has also drawn a deliberate contrast on business model. According to reports, the company has stated Claude will remain ad-free as OpenAI has been testing advertising in ChatGPT’s free tier. For enterprise buyers, that distinction matters less in absolute terms than as a signal about how each company intends to monetize attention, and whose interests it’s primarily serving.
On benchmarks: Anthropic reports that Opus 4.6 achieves the highest score on Terminal-Bench 2.0, an agentic coding evaluation, and outperforms OpenAI’s GPT-5.2 by approximately 144 Elo points on GDPval-AA, which measures economically valuable knowledge work in finance and legal domains. These are vendor-reported figures from Anthropic’s own evaluations. No independent third-party assessment of either model against these benchmarks is currently available from organizations like Epoch AI.
What to watch: Frontier’s pricing disclosure will be a defining moment. Right now, enterprise buyers can’t make a cost-equivalent comparison. Watch also for Microsoft Foundry’s positioning as a distribution channel for both products simultaneously, a dynamic that may pressure both Anthropic and OpenAI to sharpen their direct-to-enterprise value propositions rather than competing through the same cloud intermediary.
The TJS read: this isn’t a technology story about which model scores higher on benchmarks. It’s a business model story. Anthropic is betting that model quality and pricing clarity win enterprise trust. OpenAI is betting that platform lock-in, shared context, agent orchestration, workflow integration, creates switching costs that performance comparisons can’t overcome. Enterprise buyers are now choosing between those two bets, and the choice has real architecture and cost implications.