Inference
EN

Claude Fable 5 vs Gemini 4 Carbon

Change models
Claude Fable 5Anthropic
Gemini 4 CarbonGoogle
IndexAll categories, out of 100.
74.0 ±2.7
82.7 ±5.1 best
GapIndex points below the top.
−8.7
Top
CoverageIndex weight that has results.
95% best
80%
SpeedOutput tokens per second.
56/s best
31/s
EloLMArena rating and rank.
1492
1538 best
ReleasedMonth of release.
2026-06
2026-10
Input / 1MUS dollars per 1M input tokens.
$10.00
$2.00 best
Output / 1MUS dollars per 1M output tokens.
$50.00
$10.00 best
Batch output / 1MBatch price per 1M output tokens.
$25.00
$5.00 best
1K chats2K tokens in, 500 out, each.
$45.00
$9.00 best
1K documents20K tokens in, 1K out, each.
$250.00
$50.00 best
1K agent tasks200K tokens in, 20K out, each.
$3,000.00
$600.00 best
ContextMaximum tokens in one request.
1M
1.05M best
AgenticMulti-step tasks with tools.
71.4
83.2 best
CodingCode writing and repair.
74.7
80.5 best
ReasoningLogic problems and puzzles.
73.6
83.9 best
MultimodalTasks with images and text.
70.1
—
KnowledgeFacts and expert knowledge.
75.8
79.5 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
64.1
—
MathMath problems.
76.2
77.2 best

Benchmarks

24 shared
Only Claude Fable 5: 68Only Gemini 4 Carbon: 15
Claude Fable 5
Gemini 4 Carbon
Artificial Analysis Long Context ReasoningReasoning
82.3% best27 / 217
79.3%52 / 217
GPQA diamond · Epoch AIKnowledge
85.9%69 / 217
96.5% best1 / 217
OTIS Mock AIME 2024-2025Math
99.7%14 / 200
100.0% best1 / 200
Artificial Analysis Omniscience AccuracyKnowledge
65.4% best4 / 164
58.3%9 / 164
ARC-AGI-2 (semi-private)Reasoning
89.2%11 / 131
92.1% best5 / 131
ARC-AGI-1 (semi-private)Reasoning
98.5% best1 / 126
97.5%5 / 126
Vibe Code Bench v1.1Coding
90.4% best4 / 123
89.1%9 / 123
ARC-AGI-2 (public eval)Reasoning
96.4%9 / 122
97.9% best3 / 122
ARC-AGI-1 (public eval)Reasoning
97.6%17 / 116
98.5% best6 / 116
SimpleQA VerifiedKnowledge
70.7%9 / 91
77.7% best2 / 91
Chess PuzzlesReasoning
47.0%24 / 90
78.0% best2 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
87.0%12 / 90
90.5% best6 / 90
Legal Research BenchLegal
49.5%9 / 90
55.8% best1 / 90
Tax Agent BenchFinance
38.3%12 / 82
51.8% best1 / 82
Mystery Game PuzzlesReasoning
52.0%14 / 80
73.0% best5 / 80
FrontierMath-Tier-4-v2-PrivateMath
90.2%7 / 72
95.1% best4 / 72
Terminal-Bench 4.0Agentic
41.4%14 / 61
63.1% best4 / 61
Agent Arena steerabilityAgentic
7.915 / 54
19.1 best2 / 54
Agent Arena task outcomeAgentic
5.114 / 54
18.4 best4 / 54
Furniture AssemblyReasoning
35.8%23 / 37
77.5% best6 / 37