Google launches Gemini 4 Argon with a 1M-token window and $2 introductory pricing
Gemini 4 Argon, Google's new flagship, is built for multi-step work in coding, finance, legal, cybersecurity and science and tops the Vals Index. It offers a unified one-million-token input-output limit via the API, with introductory pricing of $2 per million input tokens rising to $4 at the end of November.
Google has released Gemini 4 Argon, a frontier model positioned for sustained, multi-step work rather than single-turn chat. The company highlights large engineering tasks such as refactoring a codebase or optimising a production Python project overnight, along with improved spatial reasoning for 3D scenes and games, and a dedicated robotics variant that can turn a video of a human performing a task into simulation code for a humanoid robot.
Benchmarks reported by Google include first place on the Vals Index across finance, coding, legal and tax work, 87 per cent on the Embodied Reasoning benchmark, and leading scores on private evaluations of rare-disease diagnosis and wet-lab reasoning. On safety, Argon leads the Gray Swan indirect prompt-injection benchmark with attack success reduced to 0.7 per cent, and Google says shortcut-taking in coding tests fell to 2 per cent.
The model answers more directly with less hedging and follows style instructions more closely than earlier Gemini releases. Through the Gemini API it offers what Google calls an industry-leading one-million-token unified input-output limit, and cached input is discounted by 95 per cent.
Argon is available globally via AI Studio and the Gemini platform, to Ultra subscribers in Google Antigravity, and to Ultra and Pro subscribers in the Gemini app and AI Mode in Search. Introductory pricing is $2 per million input tokens and $10 per million output tokens until 30 November, after which it doubles. Google is also offering 10,000 selected scientists six months of free access to Antigravity's highest tier.
Why it matters
Argon's launch, followed two weeks later by Carbon, shows Google iterating at a pace that compresses the window in which any lab holds a clear lead. The introductory pricing is a deliberate land-grab for agentic workloads before the price rises, and the scientist programme signals where Google believes the next proof points will come from.
Indicops view
A one-million-token unified window suits document-heavy government work, where a single case file or policy corpus can run to hundreds of pages. We caution clients to budget on the post-November price rather than the introductory one, and to measure cost per completed task, not per token, since models that finish work in fewer steps are often cheaper despite higher list prices.
More in Models
Back to the newsroomGoogle launches Gemini 4 Hydrogen, a $1.50-per-million-token model for coding and agents
Google has released Gemini 4 Hydrogen, a faster and cheaper tier of its Gemini 4 family aimed at everyday coding and agentic work. It costs $1.50 per million input tokens and $7.50 per million output tokens across a 1M-token window, and is rolling out through the Gemini API, AI Studio and consumer products.
Alibaba opens a two-week preview of Qwen4-Max with a 1M-token context, but no benchmarks or weights
Alibaba Cloud has opened a time-limited preview of Qwen4-Max, its next flagship model, through the Token Plan only. It offers a 1M-token context and up to 393,216 output tokens, but the documentation lists no pricing, benchmarks or downloadable weights, and the model retires on 25 November.
Anthropic releases Claude Fable 5.2 and Mythos 5.2, 30 per cent faster and 20 per cent cheaper
Anthropic has released Claude Fable 5.2 for general use and Mythos 5.2 for vetted cyber and life-sciences organisations. The company says output is up to 30 per cent faster than the 5.1 generation, at $8 per million input tokens and $40 per million output tokens, with lower rates of safeguard blocks.