# Gemini 4 Neon

> Gemini 4 Neon is Google's November 18, 2026 addition to the Gemini 4 family (model ID: gemini-4-neon), with a March 2026 knowledge cutoff. Within Gemini 4, Neon bridges the gap between speed-oriented Hydrogen and long-horizon Carbon: Google positions it for repository-scale coding, terminal debugging, and multi-step agent orchestration where both strong judgment and low latency matter.

Canonical: https://routemux.com/google/gemini-4-neon
Provider: Google

## Key facts

| Field | Value |
| --- | --- |
| Model ID | google/gemini-4-neon |
| Official model ID | gemini-4-neon |
| Context window | 1M |
| Max output | 64K |
| Input modalities | text, image, video, audio, file |
| Output modalities | text |
| Capabilities | Vision, Reasoning, Tool Calling, Json Mode, Coding, Audio Input, Video Input |
| Upstream release | 2026-11-18 |
| Availability | coming soon |

## Pricing

- RouteMux input: $1.5 per 1M tokens
- RouteMux output: $7.5 per 1M tokens
- Official input: $3 per 1M tokens
- Official output: $15 per 1M tokens
- Cache read: $0.075 per 1M tokens
- Saving vs official: 50%

Billing is prepaid wallet, charged per successful request only.

## API access

Base URL: https://api.routemux.com

Supported protocols:

- openai_chat
- anthropic_messages
- google_vertex_ai

Aliases that also route here: models/gemini-4-neon

## What is Gemini 4 Neon

Gemini 4 Neon is Google's November 18, 2026 addition to the Gemini 4 family (model ID: gemini-4-neon), with a March 2026 knowledge cutoff. Within Gemini 4, Neon bridges the gap between speed-oriented Hydrogen and long-horizon Carbon: Google positions it for repository-scale coding, terminal debugging, and multi-step agent orchestration where both strong judgment and low latency matter.

Beyond single-model calls, Google highlights Neon as a dedicated Orchestrator—decomposing complex engineering goals, directing parallel subagents, and verifying intermediate outputs while pairing with faster execution models such as Hydrogen. For long, tool-heavy sessions, Google notes that Neon is tuned for reliable context compaction and progress tracking, cutting down redundant file reads across multi-turn coding loops.

Specs and interface limits: a shared 1,048,576-token window and a 65,536-token single-segment output cap for both streaming and non-streaming calls (unlike Carbon's 32,768 non-streaming cap). Thinking supports low, medium, high (default), and xhigh; custom temperature, top_p, and top_k values are ignored. RouteMux exposes OpenAI Chat Completions, Anthropic Messages, and native Gemini endpoints with streaming, function tools, multimodal input, and implicit context caching from 4,096 tokens.

On RouteMux, Gemini 4 Neon is billed at $1.5 per 1M input tokens (official $3), $7.5 per 1M output tokens (official $15), and $0.075 per 1M cached input tokens (official $0.15)—saving 50% compared with Google's list price.

## Key capabilities

- **Built to act as a multi-agent Orchestrator** — Handles end-to-end repository edits standalone or serves as a planning layer that delegates parallel subtasks to Hydrogen and verifies results.
- **Long-session context compaction** — Google tunes Neon to compact long tool histories reliably and avoid redundant file reads across extended coding and terminal sessions.
- **High by default, xhigh available, 65,536 both ways** — Offers low, medium, high (default), and xhigh thinking levels, keeping a full 65,536-token segment cap whether streaming or non-streaming.
- **March 2026 cutoff with multimodal & implicit cache** — Accepts text, image, video, audio, and PDF inputs, with implicit caching kicking in at 4,096 tokens for repeated repository or document prefixes.

## Other variants in this family

- Gemini 4 Argon (`google/gemini-4-argon`)
- Gemini 4 Carbon (`google/gemini-4-carbon`)
- Gemini 4 Hydrogen (`google/gemini-4-hydrogen`)

## Sources

- [Google AI Studio: Gemini 4 Neon model documentation](https://aistudio.google.com/docs/models/gemini-4-neon)
- [Google Cloud: Gemini 4 Neon model page](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/4-neon)
- [Google: Get started and migrate to Gemini 4](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/start/get-started-with-gemini-4)
