Alphabet Inc.

(NASDAQ Global Select)
Back to all articles
Bullish +65

Alphabet (GOOGL) Stock Rises 2% Pre-Market After Gemini 4 Argon Launch

AI model release AI-generated analysis

πŸ“ˆ Alphabet (GOOGL) stock climbed about 2% in pre-market trading after Google unveiled Gemini 4 Argon.

πŸ’° Argon is priced at $2 per million input tokens and $10 per million output tokens, undercutting Anthropic's Claude Opus 5.5.

πŸš€ The model features an industry-leading 1M token output limit, a major jump from the previous 64,000 token cap.

πŸ’» Argon scored a record 77.9% on the DeepSWE v1.1 coding test and ranked first for coding on LLM Arena's Text Arena.

πŸ›‘οΈ The model demonstrated strong cybersecurity capabilities with a lower rate of failed prompt-injection attacks at 0.7% compared to competitors.

🀝 Early access is granted to cybersecurity researchers, paid API customers, and Google AI Ultra subscribers before public release.

πŸ“‰ Jefferies analysts maintain a 'buy' rating with a $445 price target, citing the need to confirm real-world utility of early benchmarks.

πŸ›οΈ CEO Sundar Pichai attended a White House meeting on AI safety and signed a statement backing voluntary internal safety controls.

πŸ” The launch addresses investor worries about Google's DeepMind lab following recent slower releases and weaker benchmark scores.

Bullish Signals
  • Alphabet (GOOGL) stock rose approximately 2% in pre-market trading following the strategic launch of Gemini 4 Argon.
  • The new model offers aggressive pricing at $2 per million input tokens, significantly cheaper than competitor Claude Opus 5.5.
  • Argon delivers an industry-leading 1M token output limit, providing a substantial capacity increase over previous models and competitors.
  • The model achieved a record-breaking 77.9% score on the DeepSWE v1.1 coding test, demonstrating superior performance in software engineering tasks.
  • Argon exhibits enhanced cybersecurity resilience with a lower prompt-injection failure rate of 0.7% compared to major rivals like Claude Opus 5.5.
  • The model ranks first for coding on LLM Arena's Text Arena, indicating strong competitive positioning in developer tools.
Risk Factors
  • Jefferies analysts noted that Argon still trails top models on the tougher FrontierSWE v2 test, suggesting room for improvement in complex reasoning.
  • The company has not specified how long the introductory pricing and 95% discount on cached tokens will remain available.
  • Real-world utility of the model's early benchmark numbers still needs confirmation before broader adoption.
  • Google plans to keep tightening safeguards before Argon reaches the public, potentially delaying full market availability.
Full Analysis
Alphabet (GOOGL) shares rose approximately 2% in pre-market trading following the launch of Gemini 4 Argon, a new AI model unveiled Wednesday afternoon. The stock had already gained roughly 1% during regular trading hours before this announcement, with earlier highs near 3%. CEO Sundar Pichai highlighted that internal teams are immediately utilizing the model for complex tasks including coding, quantum computing, and cybersecurity defense. Gemini 4 Argon is positioned as a cost-effective alternative to competitors like Anthropic's Claude Opus 5.5, with input token pricing set at $2 per million compared to $4 for the rival model. Google offers a significant 95% discount on cached input tokens and provides early access to cybersecurity researchers and paid API customers before a broader public rollout. The model features an industry-leading output limit of 1 million tokens, significantly higher than Anthropic's 128,000 token limit. Benchmark results show Argon achieved a record 77.9% score on the DeepSWE v1.1 coding test and ranked first for coding on LLM Arena's Text Arena. The model also demonstrated strong cybersecurity capabilities, tying Grok 4.7 for the top score on vulnerability fix tests and showing a lower rate of failed prompt-injection attacks compared to major competitors. Analysts from Jefferies maintain a 'buy' rating with a $445 price target, noting that real-world utility must still be confirmed despite early benchmark successes. The launch aims to address investor concerns regarding Google's DeepMind AI lab, which had faced scrutiny over slower releases and weaker scores recently. Pichai attended a White House meeting on AI safety alongside other tech executives, signaling a focus on regulatory alignment and voluntary safety controls. While OpenAI and Anthropic did not comment on the specific claims, the pricing strategy and technical benchmarks suggest a competitive push in the enterprise AI market.