Google Gemini 4 Argon Matches GPT-6 Astra on Intelligence Index at 60% of the Cost

Google has introduced Gemini 4 Argon, marking its first non-Flash-class model release in seven months. This update matters because the new architecture delivers performance that rivals the current market leader, GPT-6 Astra, while offering a significantly lower entry price for developers. We can now see a direct shift in the competitive landscape as Google pushes into the high-end coding and reasoning tier.

New model leads coding benchmarks while undercutting rivals on price

The model arrives with a 1 million token context window, allowing it to process vast amounts of data in a single pass. Google teams are already using the system extensively for internal projects, and it is currently undergoing testing by the US government. This early deployment suggests the company has high confidence in the model's stability and security for sensitive applications.

Gemini 4 Argon leads the DeepSWE v1.1 benchmark with a score of 77.9 percent, outperforming both GPT-6 Astra and Claude Opus 5.5. It also ranks first on AutomationBench-AA with a 77.5 percent score and achieves 57 percent on Terminal Bench 4. These results indicate strong capabilities in automated software engineering and complex terminal-based tasks.

On the Artificial Analysis Intelligence Index, the model scores 53, matching the maximum score of GPT-6 Astra while staying ahead of GPT-6.1 Sol. It maintains the lowest hallucination rate of 15 percent among all models scoring above 45 on this index. The model also holds the fourth position on the Vals RSI Index, trailing Claude Opus 5.5, Fable 5.1, and Opus 5.0.

Pricing for the model is currently discounted to $2 per million input tokens and $10 per million output tokens. This structure equates to approximately $210 for standard usage tiers, which is 60 percent of the cost of GPT-6 Astra (max) at the current promotional rate. Artificial Analysis notes that these prices will rise to $3.98 per Intelligence Index task once the discount period ends.

Sundar Pichai shared an early look at the model to address growing discussion about Google's next generation of AI. The release confirms Google's ability to deliver competitive capabilities, though some internal reports suggest the model still struggles with specific coding tasks. The current confirmed facts establish Gemini 4 Argon as a strong contender in the top-tier AI market.