Google opens Gemini 4 Argon to paid API users and Gemini subscribers two weeks after limited debut

The developer post moves Argon out of its restricted preview. Paid API customers and Gemini app subscribers can now call the model, where two weeks earlier only a small early-access group could. Google describes it as its new frontier model for difficult problems and says it is now broadly available to developers, enterprises, scientists and consumers. Reuters, as relayed by The Neuron, reports that Google is positioning Argon as a flagship release meant to win back ground against OpenAI and Anthropic. The company reached this point after months of delays and a leadership shakeup at DeepMind, so the broad rollout is also a test of whether the model holds up outside a curated preview. For builders, general availability is the moment to run real evaluations. Teams that route between providers can now add Argon to their eval harness, compare cost and latency on their own workloads, and set fallback order. Anyone locked to one lab for a production agent gets a third frontier option to price against, which strengthens negotiating leverage on committed spend.