Home/Models/Gemini 4 Carbon
Google DeepMind logo

Gemini 4 Carbon

by Google DeepMind · Released Nov 2026

1M Context
92.9
avg score
Rank #4
Compare
Top 1% of all models
Context
1.0M tokens (~524 books)
Input $/1M
$2.00
Output $/1M
$10.00
Type
text
License
Proprietary
Benchmarks
17 tested
Data as of
About

Gemini 4 Carbon is Google DeepMind frontier model, generally available since 2026-11-10 through the Gemini API in Google AI Studio and Gemini for business for all paying developers (limited preview to cyber defenders from 2026-10-28) · introductory price $2 input and $10 output per 1M tokens until 2026-11-30, cached input 95% off · list price from 2026-12-01 $4 input and $20 output · unified context window 1,048,576 tokens with default output 65,536 tokens.

Tested on 17 benchmarks · BenchGecko score 92.9. Top scores: OTIS Mock AIME 2024-2025 (100.0%), GPQA diamond (99.3%), ARC-AGI (98.5%).

Looking for similar performance at lower cost?
Step 3.5 Flash scores 89.5 (96% as good) at $0.10/1M input · 95% cheaper
Capabilities
coding
61.2
#41 globally
reasoning
94.6
#2 globally
math
96.8
#6 globally
knowledge
73.7
#8 globally
agentic
84.2
#1 globally
general
80.1
#4 globally
Benchmark Scores
Compare All
Tested on 17 benchmarks · Ranked across 6 categories
Score Distribution (all 329 models)
0255075100
▲ You are here
ARC-AGI

Abstraction and Reasoning Corpus. Tests fluid intelligence through novel visual pattern recognition puzzles. Core measure of general intelligence.

98.5·
ARC-AGI-2

ARC-AGI 2, harder sequel to ARC. More complex abstract reasoning patterns that test generalization ability beyond training data.

94.2·
SimpleBench

Deceptively simple questions that humans find easy but AI models often get wrong. Tests common sense and reasoning gaps.

91.1·
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

100.0·
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Recently Happened
Gemini 4 Carbon report card updated: Tokenizer Tax
Nov 16, 2026
Links
Documentation
BenchGecko API
gemini-4-carbon
Specifications
  • Typetext
  • Context1.0M tokens (~524 books)
  • ReleasedNov 2026
  • LicenseProprietary
  • StatusActive
  • Cost / Message~$0.014
Available On
Google DeepMind logoGoogle DeepMind$2.00
Share & Export
Tweet
Gemini 4 Carbon is a proprietary text AI model by Google DeepMind, released in November 2026. It has an average benchmark score of 92.9. Context window: 1M tokens.

Key facts · as of 2026-11-22

  • Gemini 4 Carbon by Google DeepMind. BenchGecko score 92.9, rank 4 of 329 scored models (normalized average of public benchmark scores).
  • List price $2.00 input · $10.00 output per 1M tokens (as of 2026-11-22).
  • Sold by 5 providers (as of 2026-11-22): Google AI Studio $1.00 in / $5.00 out · Google $2.00 in / $10.00 out · Google AI Studio $2.00 in / $10.00 out · Google $3.60 in / $18.00 out · Google AI Studio $3.60 in / $18.00 out. Every provider
  • Gecko Tests · Gecko Score 55 (rank 21): Who Are You B (Knows who made it) · Censorship Index E (Refuses or dodges 5%) · Tokenizer Tax A (39% more tokens outside English).

How to cite · data as of 2026-11-22

Gemini 4 Carbon · benchmarks, pricing and providers. BenchGecko, data as of 2026-11-22. https://benchgecko.ai/model/gemini-4-carbon

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP