Inference
EN

Claude Fable 5.1 vs Claude Sonnet 5.5

Change models
Claude Fable 5.1Anthropic
Claude Sonnet 5.5Anthropic
IndexAll categories, out of 100.
76.2 ±2.4
77.2 ±2.6 best
GapIndex points below the top.
−1.0 · no clear gap
Top
CoverageIndex weight that has results.
95%
95%
SpeedOutput tokens per second.
69/s
107/s best
EloLMArena rating and rank.
1512 best
1469
ReleasedMonth of release.
2026-09
2026-09
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
1M
AgenticMulti-step tasks with tools.
71.1
77.5 best
CodingCode writing and repair.
77.7
80.8 best
ReasoningLogic problems and puzzles.
75.8
77.8 best
MultimodalTasks with images and text.
77.9 best
74.4
KnowledgeFacts and expert knowledge.
77.2 best
71.3
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
74.3 best
41.5
MathMath problems.
77.0 best
76.1

Benchmarks

94 shared
Only Claude Fable 5.1: 38Only Claude Sonnet 5.5: 34
Claude Fable 5.1
Claude Sonnet 5.5
Artificial Analysis Humanity's Last ExamKnowledge
59.1% best3 / 222
55.0%6 / 222
Artificial Analysis Long Context ReasoningReasoning
85.3% best5 / 217
82.7%23 / 217
Critical Physics TasksReasoning
29.7%12 / 214
31.4% best7 / 214
OTIS Mock AIME 2024-2025Math
100.0%1 / 200
100.0%1 / 200
Artificial Analysis Omniscience AccuracyKnowledge
68.4% best2 / 164
51.5%17 / 164
GDPval-AA normalizedAgentic
62.7%6 / 133
66.9% best4 / 133
ARC-AGI-2 (semi-private)Reasoning
90.0% best9 / 131
83.9%17 / 131
ARC-AGI-1 (semi-private)Reasoning
97.5% best5 / 126
96.5%11 / 126
Vibe Code Bench v1.1Coding
90.3%6 / 123
92.4% best2 / 123
ARC-AGI-2 (public eval)Reasoning
98.3% best2 / 122
90.1%16 / 122
ARC-AGI-1 (public eval)Reasoning
99.0% best2 / 116
97.8%14 / 116
Artificial Analysis SciCodeCoding
63.1% best3 / 114
61.0%6 / 114
MedScribeHealthcare
91.3% best2 / 109
91.1%3 / 109
MedCodeHealthcare
53.5% best8 / 106
52.9%12 / 106
SimpleQA VerifiedKnowledge
70.8% best8 / 91
46.5%39 / 91
Chess PuzzlesReasoning
48.0%19 / 90
59.0% best10 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
90.2% best7 / 90
88.8%10 / 90
Legal Research BenchLegal
55.3% best2 / 90
48.1%12 / 90
SWE-bench ProCoding
81.2%4 / 83
81.3% best3 / 83
Tax Agent BenchFinance
50.7% best2 / 82
42.0%10 / 82