Anthropic’s Claude has emerged as the clear favorite to be rated the best artificial intelligence model by the end of 2026, according to betting data from Kalshi. The federally authorized prediction platform shows that over $7.6 million has been wagered on the outcome, with Claude commanding a 65.2% probability of winning the title. This figure represents a 2.7% increase in confidence among bettors, signaling a significant shift in market sentiment away from established leaders like OpenAI and Alphabet Inc.

OpenAI’s ChatGPT trails significantly in second place with a 14.6% probability, while Alphabet Inc.’s Google Gemini sits third at 10.2%. The disparity highlights a growing investor belief that Anthropic is gaining ground in the competitive AI landscape. The stakes are high for all three companies, as dominance in large language models drives valuation multiples and strategic partnerships across the technology sector.

Market Probabilities Comparison

The surge in Claude’s odds coincides with Anthropic’s recent launch of Opus 5, its latest artificial intelligence model. Anthropic stated that Opus 5 delivers major improvements in software engineering, business automation, and scientific research. Specifically, the company claimed that Opus 5 offers double the performance of its predecessor, Opus 4.8, on the Frontier-Bench benchmark while simultaneously lowering the cost per task.

Competitor Infrastructure Moves

While Anthropic focuses on model efficiency, competitors are scaling infrastructure aggressively. Nvidia is reportedly considering backing $250 billion in financing to support OpenAI’s massive AI infrastructure expansion. This funding would support a data center project in southern Ohio, which could cost more than $500 billion. Such capital intensity underscores the heavy investment required to maintain relevance in the generative AI race.

What the Numbers Show

The wide gap between Claude’s 65.2% probability and the combined total of its two nearest rivals (24.8%) suggests a strong consensus among prediction market participants regarding Anthropic’s current trajectory. This concentration of bets indicates that recent product launches, particularly Opus 5, have materially altered perceptions of competitive advantage in the sector.

Chamath Palihapitiya, chief of Social Capital, has issued a stark warning regarding the valuation of artificial intelligence model companies, labeling large terminal values for the sector as a "mathematical mistake." The critique arrives as two major players in the space, Anthropic and OpenAI, advance their preparations for public market listings. Palihapitiya argues that the sustainable business advantage in AI has shifted away from the foundational model layer itself, moving instead to the application layer above it and the infrastructure layer below it.

The catalyst for this reassessment appears to be the emergence of Moonshot AI’s Kimi K3, an open-weight Chinese model that reportedly matches frontier performance at a fraction of the cost. Palihapitiya noted on the All-In podcast that once a laboratory publishes its performance metrics, rivals are able to match or exceed them within weeks. He highlighted that this pace of commoditization, which historically took five to ten years in other industries, is now occurring in months, thereby eroding the long-term moat previously assumed for proprietary model developers.

Regulatory and Market Context

The debate over open-weight models has intensified at the policy level. The White House spent last week weighing restrictions on open source AI models, particularly those originating from China, following the release of Kimi K3. The industry response was swift and divided. Nvidia Corp., Meta, and 23 other companies signed an open letter defending open weights, while Google, Amazon, and closed-lab developers declined to sign. David Sacks, who stepped down as White House AI czar in March and now co-chairs the President’s Council of Advisors on Science and Technology, described a potential ban as a "tragic mistake" that would backfire on the United States.

Market sentiment reflects uncertainty around regulatory intervention. Polymarket data indicates approximately 14% odds that the US government will ban an open source AI model this year. A related contract tracking Washington removing public access to a major Chinese model sits at 11%. For Anthropic and OpenAI, a ban would arguably represent the most favorable outcome, effectively removing their lowest-cost competition by decree.

Implications for Public Listings

Both Anthropic and OpenAI are currently engaged in road shows, facing direct scrutiny from investors regarding the threat of commoditization from open-source alternatives. Anthropic filed confidentially with the Securities and Exchange Commission (SEC) on June 1, with OpenAI following one week later. Prediction markets price roughly 71% odds that Anthropic will conduct an initial public offering by the end of the year, compared to 19% odds for OpenAI within the same timeline.

Palihapitiya’s analysis suggests that investment opportunities may lie beneath the model layer rather than within it. He singled out Alphabet Inc., arguing that a landscape with hundreds of competing models benefits companies selling the silicon and cloud infrastructure required to serve them. Nvidia Corp. occupies a similar position in this trade thesis, as cheaper models could drive higher inference volumes rather than reducing demand. Palihapitiya clarified that his concern is not whether these companies will generate revenue or produce quality models, but rather that the current pricing of the model layer fails to account for its rapid erosion into a commodity.