Inference
EN

Gemini 4 Carbon vs Gemini 4 Hydrogen

Change models
Gemini 4 CarbonGoogle
Gemini 4 HydrogenGoogle
IndexAll categories, out of 100.
82.3 ±3.8 best
76.9 ±8.4
GapIndex points below the top.
Top
−5.4 · no clear gap
CoverageIndex weight that has results.
90% best
65%
SpeedOutput tokens per second.
31/s
—
EloLMArena rating and rank.
1538
—
ReleasedMonth of release.
2026-10
2026-11
Input / 1MUS dollars per 1M input tokens.
$2.00
$1.50 best
Output / 1MUS dollars per 1M output tokens.
$10.00
$7.50 best
Batch output / 1MBatch price per 1M output tokens.
$5.00
—
1K chats2K tokens in, 500 out, each.
$9.00
$6.75 best
1K documents20K tokens in, 1K out, each.
$50.00
$37.50 best
1K agent tasks200K tokens in, 20K out, each.
$600.00
$450.00 best
ContextMaximum tokens in one request.
1.05M best
1M
AgenticMulti-step tasks with tools.
82.9 best
78.5
CodingCode writing and repair.
81.1 best
79.1
ReasoningLogic problems and puzzles.
81.6
—
MultimodalTasks with images and text.
84.3 best
80.5
KnowledgeFacts and expert knowledge.
79.3 best
74.3
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
77.0
—

Benchmarks

9 shared
Only Gemini 4 Carbon: 34Only Gemini 4 Hydrogen: 2
Gemini 4 Carbon
Gemini 4 Hydrogen
Artificial Analysis Omniscience AccuracyKnowledge
58.3% best9 / 165
50.5%19 / 165
FrontierCode 2.0Coding
40.0% best5 / 28
38.2%7 / 28
Artificial Analysis AutomationBenchAgentic
77.8% best1 / 23
76.1%4 / 23
Artificial Analysis GDP.pdfAgentic
28.4% best7 / 22
27.6%8 / 22
Artificial Analysis Terminal-Bench v4.0Agentic
65.2% best2 / 22
53.5%10 / 22
cursorBench40Coding
60.4% best1 / 21
54.3%5 / 21
Multimodal Knowledge Agent BenchmarkAgentic
54.3% best2 / 20
53.4%3 / 20
OSWorld 2.1 offline subset (82 tasks)Agentic
81.5% best5 / 11
77.6%6 / 11
Chartography without toolsMultimodal
72.5% best2 / 10
66.6%4 / 10