Google Gemini 4 Neon breaks automation records as Anthropic goes public
Google's latest model hit a new high on Zapier's AutomationBench just as its rival Anthropic debuted on the stock market.
The short version
- Gemini 4 Neon reached 56.3% on AutomationBench, a test of 657 real business workflows.
- Google has set three records on the platform in the last seven weeks.
- Alphabet owns nearly 14% of Anthropic and secured $121 billion in hosting through 2033.
- Gemini 4 Neon costs $0.88 per task on high-effort tasks, less than half of its predecessor's $1.94.
- Anthropic's Claude Fable 5.2 scored 51.4% on the same automation benchmark.
Why it matters
This shows Google is successfully optimizing for cost-effective business automation while simultaneously profiting from its competitors' infrastructure needs. For investors, it highlights a unique position where Alphabet wins regardless of which AI model takes the lead.
New benchmarks for automation
Google recently launched Gemini 4 Neon, which immediately set a new high score of 56.3% on Zapier’s AutomationBench. This specific test evaluates how well an AI can handle 657 real-world business workflows. It marks the third time Google has broken a record on this platform in just seven weeks.
The progression of the Gemini 4 family has been rapid. The Argon model first passed the 50% mark on September 30, followed by the Carbon model reaching 53.9% on October 28. The current Neon model continues this upward trend, outperforming Anthropic’s best model, Claude Fable 5.2, which scored 51.4%.
Zapier CEO Wade Foster noted that while the public often focuses on the rivalry between OpenAI and Anthropic, Google has consistently taken the top spot on practical business tasks. This shift suggests that for many companies, the ability to execute complex workflows is becoming the primary metric for success.
Efficiency and coding gains
Beyond raw performance, Gemini 4 Neon is designed to be more cost-effective for high-effort tasks. The model scores 54.3% on these difficult tasks for $0.88 per task. This is less than half the cost of the previous Carbon model, which required $1.94 to achieve similar results.
The model also showed significant progress in coding, a specific area where Alphabet CEO Sundar Pichai previously identified a need for improvement. Gemini 4 Neon currently holds the number one spot on Cursorbench, a platform used to measure coding capabilities. This is a direct competitive move against Anthropic's Claude Code.
Elon Musk even commented on the release, calling the new model "a banger" in a reply to Pichai. This public recognition highlights the growing visibility of Google's iterative improvements in the AI space. The focus remains on making these models cheaper and more capable for developers and enterprise users alike.
The Anthropic IPO dynamic
The timing of these records is notable because they occurred just as Anthropic went public. Anthropic priced its initial public offering at $83 a share and closed its first session at $97.17, a 17.1% increase. This debut happened while Microsoft and Amazon were also competing for dominance in the AI market.
Alphabet occupies a unique position in this landscape because it is both a competitor and a major partner to Anthropic. Google owns nearly 14% of the company and has secured a massive commitment from Anthropic for at least $121 billion in hosting through July 2033. This means Google profits from Anthropic's growth even as they compete for the same customers.
Amazon also holds a significant stake in the rivalry, having invested $18 billion in Anthropic. However, the hosting deal gives Alphabet a more direct financial link to Anthropic's infrastructure needs. While Microsoft is pitching its Copilot as a cheaper alternative to Claude, Google is positioned to capture revenue from multiple sides of the AI trade.
Financial outlook and infrastructure
Despite the technical wins, Alphabet faces significant pressure to turn these AI capabilities into profit. The company's capital spending has nearly quadrupled in four years, and management expects it to reach between $195 billion and $205 billion by 2026. These massive investments are required to power the hardware that runs models like Gemini 4 Neon.
The company reported a negative free cash flow of $9.1 billion in the third quarter, making it essential for the models to prove their value. Analysts are watching for revenue to nearly double between 2025 and 2028. To justify the spending, Google needs to maintain a growing cloud backlog and keep processing tokens at high volumes.
Currently, Alphabet trades at about 20x forward earnings, which is lower than its five-year average of 23x. This suggests that the market is waiting to see if the company can successfully monetize its AI infrastructure. The next major test for the company's valuation will be the fourth-quarter report arriving in early February.
What to watch
- Alphabet's fourth-quarter report in early February to see if revenue growth justifies capital spending.
- How many tokens Google processes per minute as they scale their cloud infrastructure.
Try it
Check the current scores on Zapier's AutomationBench to see how different models compare on real-world business workflows.
Sources and links
- Postweighed in x.com
- Postput it on X x.com
- ReportingTIKR.com: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by TIKR.com · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
