Google DeepMind Launches Gemini 4 Neon: Frontier Intelligence for Coding, Agents, and Cyber Defense
Google DeepMind launches Gemini 4 Neon between Carbon and Hydrogen, scoring 61 on the Artificial Analysis Index and 90.5% on SEC-bench Pro with $3/$15 per million token pricing.
Google DeepMind has officially released Gemini 4 Neon, a new mid-tier flagship model positioned between Gemini 4 Carbon and Gemini 4 Hydrogen. Delivering frontier-level reasoning with substantially faster end-to-end completion times, Neon scores 61 on the Artificial Analysis Intelligence Index, leads multiple cybersecurity benchmarks, and rolls out immediately across Google AI Studio, the Gemini API, Google Antigravity, and the Gemini app for Pro and Ultra subscribers.
Matching Carbon Intelligence in Less Than Half the Time
Adapted from Gemini 4 Carbon, Gemini 4 Neon supports text, image, audio, video, and PDF inputs across a unified 1,048,576-token context window with four tunable thinking levels (low, medium, high, and xhigh). On the Artificial Analysis Intelligence Index, Neon reaches a score of 61 at xhigh thinking—the highest score of any Gemini model to date—and 60 at its default high setting, matching Carbon high score while finishing in 14.8 minutes instead of 36.1 minutes. It also ranks second on the Vals Index and second on steerability on the Agent Arena Index.
State-of-the-Art Cyber Defense and Antigravity Orchestration
Google also detailed a specialized cyber defense variant coming to Fairwind Program partners. Gemini 4 Neon for cyber achieved state-of-the-art marks across SEC-bench Pro (90.5%), WebExploitBench v1.2 (51.26%), and AutoBaxBench hard challenges (76%), while reaching 70.7% on Collinear CWE-bench v2 for automated vulnerability patching. Inside Google Antigravity, Neon serves as the new default model and powers Orchestrator mode to coordinate parallel Hydrogen subagents.
API Pricing Structure and Gemini App Lineup Shift
On the Gemini API, Gemini 4 Neon carries no free tier and is priced at $3.00 per million input tokens ($0.15 per million cached input tokens) and $15.00 per million output tokens, with Batch and Flex tiers discounted by 50 percent to $1.50 and $7.50. Alongside the launch, Google streamlined its consumer Gemini app lineup to Hydrogen, Neon, and Carbon—replacing Gemini 4 Argon for Google AI Pro subscribers while reserving Carbon for Ultra tiers.