Grok 4.6 vs Hy4 preview

Change models
Hy4 previewTencent
IndexAll categories, out of 100.
66.6 ±3.7 best
66.2 ±4.0
GapIndex points below the top.
Top
−0.4 · no clear gap
CoverageIndex weight that has results.
85%
90% best
SpeedOutput tokens per second.
66/s best
60/s
EloLMArena rating and rank.
1430
—
ReleasedMonth of release.
2026-08
2026-08
Input / 1MUS dollars per 1M input tokens.
$2.00
$0.834 best
Output / 1MUS dollars per 1M output tokens.
$6.00
$2.50 best
Batch output / 1MBatch price per 1M output tokens.
$4.80
—
1K chats2K tokens in, 500 out, each.
$7.00
$2.92 best
1K documents20K tokens in, 1K out, each.
$46.00
$19.18 best
1K agent tasks200K tokens in, 20K out, each.
$520.00
$216.82 best
ContextMaximum tokens in one request.
500K
1.05M best
AgenticMulti-step tasks with tools.
67.2 best
63.1
CodingCode writing and repair.
66.6 best
64.9
ReasoningLogic problems and puzzles.
64.8
67.9 best
MultimodalTasks with images and text.
—
74.2
KnowledgeFacts and expert knowledge.
64.8
67.0 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
70.6
—
MathMath problems.
64.3
75.7 best

Benchmarks

26 shared
Only Grok 4.6: 58Only Hy4 preview: 28
Grok 4.6
Hy4 preview
17.1% best41 / 207
16.9%42 / 207
86.3% best14 / 149
83.8%47 / 149
76.2%47 / 130
77.5% best43 / 130
MedScribeHealthcare
86.5% best23 / 111
83.6%47 / 111
MedCodeHealthcare
44.7% best40 / 108
43.2%48 / 108
48.1% best15 / 98
45.2%25 / 98
33.7% best21 / 89
32.6%24 / 89
53.7%32 / 79
55.1% best23 / 79
44.6%20 / 77
47.4% best14 / 77
EMBFinance
62.7% best31 / 75
58.1%38 / 75
78.3% best13 / 75
55.1%49 / 75
17.2% best42 / 68
8.1%54 / 68
1.0% best16 / 66
0.0%35 / 66
DeepSWEAgentic
65.9% best35 / 64
64.3%39 / 64
−2.536 / 59
−1.1 best31 / 59
−2.541 / 59
3.2 best23 / 59
51.0%32 / 55
75.0% best19 / 55
19.7% best33 / 52
19.1%34 / 52
5.715 / 50
6.5 best12 / 50
66.8%17 / 50
68.6% best9 / 50