Essential Insights
- One model, many formats: Neon handles spreadsheets, PDFs, charts, audio, and video in a single 1-million-token context.
- Strong enterprise results: It ranks highly across finance, legal, tax, spreadsheets, and automation—including #1 on Legal Research, Tax Agent, and AutomationBench.
- Built for real-world agents: Neon ranks #2 on Arena’s Agent Arena Index and #1 on Vending-Bench 2.
- Excels at science and math: It leads Terminal-Bench Science, RiemannBench, and HLE-Diamond.
One model, many work formats
Gemini 4 Neon handles spreadsheets, PDFs, charts, audio and video in a single context window of up to 1 million tokens. It can combine those inputs and turn them into structured deliverables. That could help teams connect details scattered across reports and recordings. Still, organizations should test its results against their own needs before relying on them.
Strong scores across business tasks
Neon ranks No. 2 on the Vals Index, with a score of 70.9%. It ranks No. 1 on the Legal Research Bench and Tax Agent Bench. It also leads most Surge AI benchmarks in finance, accounting, spreadsheet modeling and complex instruction following. For chart reading, it scores 73.5% on Chartography. On Zapier’s AutomationBench, it ranks No. 1 with 56.3%. These results suggest broad ability, though benchmarks cannot fully predict performance in everyday work.
Promise in science, with adoption questions
Neon scores 70.5% on Terminal-Bench Science, 84.0% on RiemannBench and 71.3% on HLE-Diamond. It also ranks No. 2 on Arena’s Agent Arena Index, based on more than 4,200 real-world agent sessions. Those results point to strength in technical and scientific tasks. However, teams should weigh accuracy, cost and oversight needs before adopting it. A high score is a useful signal, not a substitute for review.
Expand Your Tech Knowledge
Explore the future of technology with our detailed insights on Artificial Intelligence.
Discover archived knowledge and digital history on the Internet Archive.
AITechV1
