gemini-4-carbon — API, Pricing & Context Window | Vivgrid
gemini-4-carbon on Vivgrid: Google's frontier Gemini 4 model for deep work on complex, real-world tasks, with a 1M-token context window, 64K max output, and text, image, video, audio, and PDF input.
gemini-4-carbon is Google's frontier model in the Gemini 4 series, built for deep work on complex, real-world tasks: long-horizon agentic coding, multi-step research, and analysis across large codebases and document sets. It pairs a ~1.05M-token context window with up to 64K max output tokens and accepts text, image, video, audio, and PDF inputs.
On Vivgrid, gemini-4-carbon runs as a globally centralized model reachable through the same OpenAI-compatible endpoint and unified key as every other model in the catalog.
Specifications
| Provider | |
| Model ID | gemini-4-carbon |
| Best for | Coding |
| Available on | Chat Completions, Vibe coding |
| Context window | 1,048,576 tokens |
| Max output | 65,536 tokens |
| Modalities | Text, Image, Video, Audio, PDF |
| Tool / function calling | Yes |
| Knowledge cutoff | 2026-03 |
| Acceleration | 🌐 Global (Centralized) |
Pricing
Pricing in USD per 1M tokens.
| Input | Cached input | Output |
|---|---|---|
| $4.00 | $0.20 | $20.00 |
What one request costs
The rates above applied to a 20K-token prompt with a 2K-token reply — a typical single agent turn, small enough that every request is billed at the base rate.
| Request | Cost |
|---|---|
| Cold prompt, nothing cached | $0.1200 |
| Warm prompt, 90% of the input served from cache | $0.0516 |
Input is priced at $4.00 per 1M tokens (one rate for text, image, video, and audio), cached input at $0.20 per 1M tokens, and output (including thinking tokens) at $20.00 per 1M tokens, with a single flat rate across the full context window — the same rates as gemini-4-argon.
Use gemini-4-carbon in your coding agent
gemini-4-carbon works with the same Vivgrid key and https://api.vivgrid.com/v1 endpoint as every other model in the catalog. Grab your API key from the Vivgrid Console, then drop one of the configurations below into your agent of choice.
OpenCode
OpenCode has Vivgrid built in, so you can use gemini-4-carbon without writing any config: open /models, search for vivgrid, paste your API key and pick the model. The OpenCode tutorial walks through it step by step.
Prefer to configure it by hand? Install OpenCode:
npm i -g opencode-aiDeclare gemini-4-carbon under the Vivgrid provider in your global OpenCode config, ~/.config/opencode/opencode.json. The @ai-sdk/openai-compatible provider talks to Vivgrid over Chat Completions (/v1/chat/completions):
{
"$schema": "https://opencode.ai/config.json",
"model": "vivgrid/gemini-4-carbon",
"provider": {
"vivgrid": {
"npm": "@ai-sdk/openai-compatible",
"name": "Vivgrid",
"options": {
"baseURL": "https://api.vivgrid.com/v1",
"apiKey": "{env:VIVGRID_API_KEY}"
},
"models": {
"gemini-4-carbon": {
"name": "gemini-4-carbon",
"tool_call": true,
"attachment": true,
"modalities": { "input": ["text", "image", "video", "audio", "pdf"], "output": ["text"] },
"limit": { "context": 1048576, "output": 65536 }
}
}
}
}
}Export your key and launch OpenCode:
export VIVGRID_API_KEY="viv-xxxxxxxxxxxxx"
opencodeUse /models to switch between gemini-4-carbon and other Vivgrid models mid-session. See the full OpenCode tutorial.
Pi
Install Pi:
npm install -g --ignore-scripts @earendil-works/pi-coding-agentAdd gemini-4-carbon to the Vivgrid provider in ~/.pi/agent/models.json. Pi talks to Vivgrid over Chat Completions (/v1/chat/completions):
{
"providers": {
"vivgrid": {
"baseUrl": "https://api.vivgrid.com/v1",
"apiKey": "viv-xxxxxxxxxxxxx",
"api": "openai-completions",
"compat": {
"sendSessionAffinityHeaders": true
},
"models": [
{
"id": "gemini-4-carbon",
"input": ["text", "image"],
"contextWindow": 1048576,
"maxTokens": 65536
}
]
}
}
}Make it the model Pi starts with in ~/.pi/agent/settings.json:
{
"defaultProvider": "vivgrid",
"defaultModel": "gemini-4-carbon"
}Then verify and start Pi:
pi --list-models vivgrid
piOr run a one-shot prompt without entering the TUI:
pi -p "explain the auth flow in this repo" --model vivgrid/gemini-4-carbonSee the full Pi tutorial.
API surfaces
The Vivgrid Console catalogs gemini-4-carbon on 2 surfaces, all behind the same API key. The endpoint shown is the one the Console previews for that surface.
| Surface | Primary endpoint | Use it for |
|---|---|---|
| Chat Completions | /chat/completions | Agent projects, where Vivgrid injects the model, system prompt and tools server-side, and Model API calls that name the model per request. |
| Vibe coding | /chat/completions | Coding CLIs and IDE agents that drive the model themselves — point the tool at Vivgrid and keep your own loop. |
Quick start
Get an API key from the Vivgrid Console, then call gemini-4-carbon directly.
curl https://api.vivgrid.com/v1/chat/completions \
-H "Authorization: Bearer $VIVGRID_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-4-carbon",
"messages": [
{ "role": "user", "content": "Say hello in English, Chinese and Spanish." }
],
"stream": true
}'Choosing within the Gemini family
Prices are USD per 1M tokens at the base rate. Rows marked this page are the model described here.
| Model | Context | Max output | Input | Output |
|---|---|---|---|---|
gemini-4-carbon — this page | 1,048,576 | 65,536 | $4.00 | $20.00 |
gemini-4-argon | 1,048,576 | 1,048,576 | $4.00 | $20.00 |
gemini-4-neon | 1,048,576 | 65,536 | $3.00 | $15.00 |
gemini-4-hydrogen | 1,048,576 | 65,536 | $1.50 | $7.50 |
gemini-3.8-flash | 1,000,000 | 128,000 | $0.75 | $3.75 |
Ideal use cases
- Deep work on complex, real-world tasks that need sustained multi-step reasoning
- Long-horizon autonomous coding agents working across large codebases
- Multimodal analysis of documents, images, video, and audio up to ~1M tokens
- Research and planning workflows where answer quality matters more than latency
Related models
- gemini-4-neon — lower-latency, lower-cost Gemini 4 model adapted from Carbon
- gemini-4-hydrogen — the fastest, lowest-cost Gemini 4 model
- gemini-4-argon — Google's flagship Gemini 4 reasoning model at the same price
deepseek-v4-pro
deepseek-v4-pro on Vivgrid: DeepSeek's flagship coding model with a 1M-token context window, up to 384K output tokens, and competitive pricing.
gemini-4-neonNew
gemini-4-neon on Vivgrid: Google's low-latency Gemini 4 model for coding, multi-step reasoning, and agentic workflows, with a 1M-token context window, 64K max output, and text, image, video, audio, and PDF input.