Google's latest model has claimed the top spot on the Vals Index, a benchmark testing AI on finance, coding, legal and tax work. Built for deep, multi-step tasks, it targets developers and enterprises, and its launch pricing could reshape the API market.
Google has released Gemini 4 Argon, which ranks #1 on the Vals Index, a benchmark of model performance across finance, coding, legal, and tax work. For developers and enterprises, it is built for deep work on complex, multi-step tasks in coding, finance, legal, cybersecurity, and science.
Key Features at a Glance
Vals Index ranking: #1 – measures performance across critical economic sectors like finance, coding, legal, and tax work.
Introductory pricing: $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off – a launch-day price for API users.
Agentic code submissions: Up to 35%+ more, with code rollbacks reduced by nearly a third – Argon in Antigravity's results for Google's engineering teams.
Gemma inference speed: 2.85x end-to-end speedup on an A100 via vLLM, with throughput rising from ~37,900 to over 108,000 tokens per second on Gemma 4 E2B, at 100% token parity – faster serving for open models.
xTB quantum chemistry: Up to 2.3x faster at predicting molecular structures and 1.8x faster at conformer search – quicker screening for drug and materials candidates.
Quantum arcsin circuit: Spacetime cost cut by more than 100x versus a published quantum CORDIC baseline – a step toward a standard library of optimized quantum circuits.
Availability
Gemini 4 Argon is available today via the API in AI Studio and the Gemini platform for developers and enterprises, for Google AI Ultra subscribers in Google Antigravity, and for Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search; Google is also offering 10,000 scientists the highest usage tier in Google Antigravity free for six months, subject to eligibility, capacity, and selection.
