Alphabet (GOOGL) Shares Jump 2% as Google Unveils Gemini 4 Argon AI Model

TLDR

• Alphabet shares advanced approximately 2% during Thursday’s premarket session.
• Google introduced Gemini 4 Argon, its latest artificial intelligence model, on Wednesday.
• The new AI system costs $2 per million input tokens and $10 per million output tokens, undercutting competitors significantly.
• Cybersecurity professionals receive priority access before wider availability to developers and general users.
• Wednesday’s regular session saw GOOGL gain around 1%, with an additional 2% climb in after-hours trading.

Alphabet (GOOGL) stock experienced a notable uptick of approximately 2% during Thursday’s premarket session. The momentum followed Google’s Wednesday afternoon introduction of Gemini 4 Argon, its cutting-edge AI platform.

Alphabet Inc., GOOGL

The Wednesday announcement occurred after shares had already posted modest gains during standard market hours. GOOGL closed regular trading up roughly 1%, following an intraday peak approaching 3% earlier in the session.

Sundar Pichai, serving as CEO, took to social platforms to highlight Argon’s capabilities. His message emphasized how Google teams are currently deploying the technology across diverse applications, from software development to advanced quantum computing research.

The rollout strategy for Argon follows a measured approach. Rather than an immediate public launch, Google has prioritized cybersecurity experts and paying API clients for initial access.

Broader availability targeting software developers, enterprise customers, and consumer markets will follow in subsequent phases. Subscribers to Google AI Ultra are included in the initial deployment wave.

Aggressive Pricing Strategy Challenges Competitors

The competitive pricing structure represents a significant element of Argon’s market positioning. Google established rates at $2 per million input tokens alongside $10 per million output tokens.

This pricing substantially undercuts Anthropic’s Claude Opus 5.5, which charges $4 and $20 respectively for equivalent token volumes. Claude Fable 5.1 carries even steeper costs at $10 and $50.

Additionally, Google is providing a substantial 95% price reduction for cached input tokens. The duration of this promotional pricing structure remains unspecified by the company.

Regarding performance metrics, Google reports Argon achieved 77.9% on the DeepSWE v1.1 coding evaluation, establishing a new benchmark. However, Jefferies analysts observed it lags behind leading alternatives on the more demanding FrontierSWE v2 assessment.

Argon secured the top position for coding capabilities on LLM Arena’s Text Arena leaderboard. On Code Arena: Web Dev, it claimed eighth place, representing a significant 21-position improvement over its predecessor, Gemini 3.8 Flash.

Security Applications Drive Development Focus

Google engineered Argon with particular emphasis on cybersecurity applications. According to the company, the platform possesses autonomous capabilities to identify, validate, and remediate software security weaknesses.

Jefferies research indicated Argon matched Grok 4.7’s leading score of 68% on CWE-bench v1, which evaluates vulnerability remediation effectiveness. Cybersecurity provider Wiz has already integrated Argon into its Scan for Good initiative.

Google documented a reduced failure rate for prompt-injection attacks at 0.7% for Argon. This compares favorably against the 1% rate observed in both Claude Opus 5.5 and Fable 5.1, according to Jefferies analysis.

The system’s output capacity has expanded dramatically. Argon now supports up to 1 million tokens, a substantial increase from the previous 64,000 limit, and significantly exceeding Anthropic’s 128,000 threshold.

Jefferies maintains a “buy” rating on Alphabet with a $445 price objective. The firm highlighted that shares had declined 16% from their May 18 high point amid concerns regarding Google’s competitive position in artificial intelligence.

The research firm emphasized that practical deployment will ultimately validate these preliminary benchmark results. Neither OpenAI nor Anthropic provided responses to inquiries regarding Argon’s performance assertions.

Wednesday’s unveiling arrives after a challenging period for Google’s DeepMind AI division. Delayed model launches, departures of prominent researchers, and disappointing benchmark performances had intensified investor apprehension prior to this week’s announcement.

Pichai participated in a White House conference addressing AI safety protocols on the day of the launch. He joined fellow technology leaders in endorsing a commitment supporting voluntary internal safety frameworks and independent audits of AI platforms.

Google indicated its intention to strengthen protective measures before expanding Argon’s availability to broader audiences. The company will conduct ongoing monitoring for potential misuse throughout the current restricted access period.