❯ Google rolls out Gemini 4 Neon, scoring 61 at its highest reasoning setting and cutting default task costs 32% versus Carbon in its comparison
Three jobs in the lineupGoogle started rolling out Gemini 4 Neon on November 18. The DeepMind model card places Neon between Carbon for difficult long-running tasks and Hydrogen for fast execution. Neon accepts text, images, audio and video, with a shared input and output context of up to one million tokens.
Scores and costs use different settingsArtificial Analysis testing gave Neon a score of 61 at its highest reasoning level. Google described that as third by model at launch, behind GPT-6.2 Astra and Claude Fable 5.2. A separate Google comparison used the default High setting and found a 32% lower cost per task than Carbon on the benchmark used. The benchmark score and the savings come from different settings.
AvailabilityGoogle’s announcement lists the Gemini API in AI Studio and Gemini for business. Gemini and Antigravity are rolling it out to Google AI Pro and Ultra subscribers. In Antigravity, Neon can also coordinate Hydrogen agents working in parallel.
▮ SIGNALNeon gives teams another way to tune reasoning depth, using the default setting to control the cost of frequent jobs while retaining strong coding performance.