Google DeepMind launches Gemini 4 Argon globally for coding, science and cyber defense workloads
Google on October 13, 2026 made Gemini 4 Argon available worldwide via the Gemini API, the Gemini app and Antigravity. The frontier model ships with a 1M-token unified context window and introductory API pricing of $2 per million input tokens and $10 per million output tokens.
Google DeepMind publicly launched Gemini 4 Argon, its latest frontier model, on October 13, 2026. The model is rolling out globally for developers and enterprises through the Gemini API, AI Studio and the Gemini platform; to Google AI Ultra subscribers in Google Antigravity; and to Ultra and Pro subscribers in the Gemini app and AI Mode in Search.
Google says Argon is built for deep, multi-step work across coding, finance, legal, cybersecurity and science, and ranks first on the Vals Index, which scores model performance on economically relevant tasks. According to The Verge, Google claims Argon outperforms rivals including GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5 on a range of benchmarks.
Pricing and availability
- Introductory API price: $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off.
- After November 30, 2026, list prices rise to $4 / $20 per million tokens.
- The Verge notes the introductory rate is roughly one-fifth the per-token cost of OpenAI’s and Anthropic’s frontier models, though total cost still depends on how many tokens a task consumes.
- The Gemini API offers an industry-leading 1M unified input-output token limit plus a new continuation protocol for long-horizon jobs.
- Google is expanding its Fairwind program for trusted cyber defenders and offering 10,000 scientists six months of Antigravity’s highest usage tier free of charge, subject to eligibility.
Internal impact and demos
Google says it has already used Argon on hard internal engineering work. In Antigravity, Argon helped raise agentic code submissions by more than 35% while cutting rollbacks by nearly a third. Cited examples include refactoring a ~6-million-line legacy folder inside Chromium’s ~40-million-line codebase, cutting ADK in-memory session prep time by 40%, and delivering a 2.85x end-to-end inference speedup for Gemma 4 E2B on vLLM.
In science, Google points to work on quantum error-correction kernels, new stellarator magnetic-field counterexamples, rare-disease genome analysis and enzyme design. In robotics, an internal “Video-to-Sim-to-Real” demo had Argon watch a human play cornhole, build a simulation, and transfer the skill to Apptronik’s Apollo humanoid, which then sank seven consecutive shots.
Safety stack
Under Google’s Frontier Safety Framework, Argon ships with stronger jailbreak resistance, a real-time misalignment monitor that reads agent reasoning and actions, and improved prompt-injection defenses. Google says unauthorized shortcut attempts fell to 2% on coding benchmarks with no incentive to hold back, and attack success on the Gray Swan Indirect Prompt Injection benchmark fell to 0.7% across 15 adaptive attempts.
Why it matters
Argon is Google’s first major frontier release since Gemini 3 in late 2025, arriving a day after Meta’s Muse Wonder and amid intense competition from OpenAI and Anthropic. Competitive pricing, a 1M-token window and tight integration with Google’s agent tooling mark a bid to re-enter the frontier race at scale.
Sources: Google Blog; The Verge.
Related articles
Tiny chip stores memories in flowing water without conventional electronics
Yale researchers have found a way to make flowing fluids remember what happened before, using...
Oxygen-based chipmaking method controls 2D crystal growth with over 99% success
Researchers in South Korea have developed a semiconductor manufacturing technique that controls where individual crystals...
Meta releases Muse Wonder, its first frontier model for personal and professional agents
Meta’s Superintelligence Lab launched Muse Wonder and Muse Wonder Full, its first frontier-class models, emphasizing Active Memory and long-running agent work. Wonder is available via the Meta Model API and Muse Code; Wonder Full is limited to Muse Code subscribers.
Anthropic to share first Q3 figures with investors as it prepares for a possible record IPO
CNBC reports Anthropic will disclose third-quarter financials for the first time to institutional investors meeting at its San Francisco headquarters this week, as the AI lab prepares for what could be a roughly $2 trillion IPO.
Reuters: OpenAI and Anthropic sought control over what outside AI safety evaluators can publish
According to Reuters, OpenAI and Anthropic pushed for significant control over outside evaluators’ publications over trade-secret concerns, even after publicly endorsing editorial independence. Talks have shifted in recent days, and both labs dispute parts of the characterization.
OpenAI and Anthropic tell UK lawmakers voluntary AI pledges are not enough
At a UK parliamentary hearing, OpenAI and Anthropic backed a statutory AI safety framework, saying voluntary commitments are not a durable plan for frontier risk. Google DeepMind and Meta declined to support new legislation, deepening a public split among labs.