Inference
EN

Claude Fable 5.2 vs Hy4

Change models
Claude Fable 5.2AnthropicSmartest
Hy4Tencent
IndexAll categories, out of 100.
85.6 ±3.7 best
72.0 ±7.8
GapIndex points below the top.
Top
−13.6
CoverageIndex weight that has results.
90% best
65%
SpeedOutput tokens per second.
64/s best
52/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-11
2026-11
Input / 1MUS dollars per 1M input tokens.
$8.00
$0.417 best
Output / 1MUS dollars per 1M output tokens.
$40.00
$1.25 best
Batch output / 1MBatch price per 1M output tokens.
$20.00
—
1K chats2K tokens in, 500 out, each.
$36.00
$1.46 best
1K documents20K tokens in, 1K out, each.
$200.00
$9.59 best
1K agent tasks200K tokens in, 20K out, each.
$2,400.00
$108.41 best
ContextMaximum tokens in one request.
1M
1.05M best
AgenticMulti-step tasks with tools.
82.2 best
74.5
CodingCode writing and repair.
86.2 best
68.4
ReasoningLogic problems and puzzles.
86.5
—
MultimodalTasks with images and text.
82.0 best
57.4
KnowledgeFacts and expert knowledge.
87.1 best
83.7
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
79.3
—

Benchmarks

13 shared
Only Claude Fable 5.2: 58Only Hy4: 8
Claude Fable 5.2
Hy4
Vibe Code Bench v1.1Coding
93.8% best1 / 123
81.0%30 / 123
Legal Research BenchLegal
46.2% best19 / 90
41.3%30 / 90
Tax Agent BenchFinance
32.1%23 / 82
34.4% best17 / 82
Humanity's Last ExamKnowledge
74.2% best1 / 77
59.1%7 / 77
Terminal-Bench 4.0Agentic
65.2% best1 / 61
39.4%15 / 61
DeepSWEAgentic
74.3%9 / 59
74.9% best8 / 59
Agent Arena steerabilityAgentic
33.0 best1 / 54
8.814 / 54
Agent Arena task outcomeAgentic
22.0 best1 / 54
6.712 / 54
HLAB-Vals-1.1Legal
5.0%27 / 35
12.5% best17 / 35
AutomationBenchAgentic
51.4%11 / 32
56.2% best4 / 32
FrontierCode 2.0Coding
49.1% best1 / 27
31.0%9 / 27
ProgramBench: Can Language Models Rebuild Programs From Scratch?Coding
85.6% best4 / 19
30.0%12 / 19
Terminal-Bench-Science 0.1Agentic
74.6% best2 / 11
19.8%11 / 11