SemiAnalysis benchmark results place Alphabet Inc.'s Google close to Anthropic on AI performance at roughly one-third the cost. Google's Gemini 4 Carbon ranked below Anthropic's flagship Claude Opus 5.5 but above OpenAI's GPT-6 Astra. SemiAnalysis's GDPx benchmark gave Carbon a score of 10.9%, against 12% for Opus 5.5. Under Google's introductory pricing, each Carbon task cost $16.29, compared with $51.41 for Anthropic. Google's earlier Gemini 4 Argon scored 7.3%.
SemiAnalysis researchers wrote on Monday that Carbon substantially improved Google's performance and was the only model broadly comparable to Opus and Astra. Carbon achieved the highest results in software engineering and biology. However, SemiAnalysis researchers said weaker safety restrictions probably contributed to Carbon's advantage in biology.
SemiAnalysis created GDPx to challenge leading models as existing benchmarks become easier for those models. SemiAnalysis collaborated with 14 specialist partners to create 206 difficult tasks covering 11 fields. The fields included finance, semiconductor design and scientific research. The highest score was only 12%, suggesting GDPx can continue measuring improvements in future models.
Polymarket traders favor Google to have the world's leading AI model at the end of 2026, despite Anthropic's higher GDPx score. Google's probability is roughly 57% in a market that has attracted more than $3.8 million in trading. Anthropic's probability is 34%. OpenAI's probability is 5%. The market uses the Arena leaderboard to determine the outcome, not GDPx.
Google forecasts capital expenditures between $195 billion and $205 billion this year. The spending increases pressure on Google to demonstrate returns from AI investment. Worries about the spending helped drive a recent decline in Google's shares. Google Cloud's third-quarter revenue rose 103% to $30.7 billion.
Carbon's combination of near-Anthropic performance and lower pricing could make Google more attractive to paying AI customers. Carbon remains unavailable to many users, and the introductory pricing could change. Carbon's initial results suggest Google's substantial investment is helping Google develop competitive models that could capture more of the AI market.
Google AI Closes Gap with Anthropic at One-Third the Cost
BenzingaShare
Summary
Google AI closes gap with Anthropic at one-third the cost, showing strong performance in software engineering and biology. On-chain data from SemiAnalysis shows Gemini 4 Carbon scored 10.9% on GDPx, versus 12% for Claude Opus 5.5, at $16.29 vs $51.41 per task. Polymarket gives Google a 57% chance of leading by 2026. Altcoins to watch may react to AI cost shifts.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.