Inference
EN

Claude Fable 5.2 vs Gemini 4 Hydrogen

Change models
Claude Fable 5.2AnthropicSmartest
Gemini 4 HydrogenGoogle
IndexAll categories, out of 100.
85.7 ±3.6 best
76.9 ±8.4
GapIndex points below the top.
Top
−8.8 · no clear gap
CoverageIndex weight that has results.
90% best
65%
SpeedOutput tokens per second.
64/s
—
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-11
2026-11
Input / 1MUS dollars per 1M input tokens.
$8.00
$1.50 best
Output / 1MUS dollars per 1M output tokens.
$40.00
$7.50 best
Batch output / 1MBatch price per 1M output tokens.
$20.00
—
1K chats2K tokens in, 500 out, each.
$36.00
$6.75 best
1K documents20K tokens in, 1K out, each.
$200.00
$37.50 best
1K agent tasks200K tokens in, 20K out, each.
$2,400.00
$450.00 best
ContextMaximum tokens in one request.
1M
1M
AgenticMulti-step tasks with tools.
83.0 best
78.5
CodingCode writing and repair.
85.8 best
79.1
ReasoningLogic problems and puzzles.
86.7
—
MultimodalTasks with images and text.
83.4 best
80.5
KnowledgeFacts and expert knowledge.
87.0 best
74.3
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
79.1
—

Benchmarks

11 shared
Only Claude Fable 5.2: 62
Claude Fable 5.2
Gemini 4 Hydrogen
Artificial Analysis Omniscience AccuracyKnowledge
79.1% best1 / 165
50.5%19 / 165
FrontierCode 2.0Coding
49.1% best1 / 28
38.2%7 / 28
Artificial Analysis AutomationBenchAgentic
76.4% best3 / 23
76.1%4 / 23
Artificial Analysis GDP.pdfAgentic
24.2%13 / 22
27.6% best8 / 22
Artificial Analysis Terminal-Bench v4.0Agentic
59.6% best4 / 22
53.5%10 / 22
cursorBench40Coding
60.0% best2 / 21
54.3%5 / 21
Multimodal Knowledge Agent BenchmarkAgentic
56.6% best1 / 20
53.4%3 / 20
CWE-bench v2Agentic
75.0% best1 / 14
68.1%2 / 14
Terminal-Bench-Science 0.1Agentic
74.6% best2 / 12
51.4%9 / 12
OSWorld 2.1 offline subset (82 tasks)Agentic
89.3% best1 / 11
77.6%6 / 11
Chartography without toolsMultimodal
73.4% best1 / 10
66.6%4 / 10