Kimi K2.5 vs Qwen3.6-27B

Change models
Kimi K2.5Moonshot AI
IndexAll categories, out of 100.
51.6 ±2.2 best
51.5 ±2.4
GapIndex points below the top.
Top
−0.1 · no clear gap
CoverageIndex weight that has results.
100%
100%
SpeedOutput tokens per second.
22/s
44/s best
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-01
2026-04
Input / 1MUS dollars per 1M input tokens.
$0.50
$0.32 best
Output / 1MUS dollars per 1M output tokens.
$2.50 best
$3.20
1K chats2K tokens in, 500 out, each.
$2.25
$2.24 best
1K documents20K tokens in, 1K out, each.
$12.50
$9.60 best
1K agent tasks200K tokens in, 20K out, each.
$150.00
$128.00 best
ContextMaximum tokens in one request.
262K
262K
AgenticMulti-step tasks with tools.
48.9
55.9 best
CodingCode writing and repair.
53.2 best
52.2
ReasoningLogic problems and puzzles.
51.1 best
48.6
MultimodalTasks with images and text.
51.6
52.6 best
KnowledgeFacts and expert knowledge.
53.2 best
48.0
MultilingualTasks in many languages.
52.1
60.7 best
InstructionTasks with strict prompt rules.
52.8 best
43.3
MathMath problems.
53.1 best
50.3

Benchmarks

42 shared
Only Kimi K2.5: 47Only Qwen3.6-27B: 22
Kimi K2.5
Qwen3.6-27B
87.6% best65 / 220
84.8%75 / 220
30.7% best73 / 216
23.1%100 / 216
78.0% best70 / 211
77.3%76 / 211
3.1% best93 / 207
1.1%113 / 207
92.2% best57 / 203
66.7%109 / 203
87.9% best51 / 193
84.2%80 / 193
35.2% best68 / 170
19.6%136 / 170
87.6%43 / 144
87.8% best42 / 144
95.9% best13 / 143
94.2%24 / 143
70.2% best42 / 140
67.6%50 / 140
17.0%83 / 132
24.3% best75 / 132
76.8%39 / 114
77.2% best37 / 114
75.4% best48 / 109
74.6%55 / 109
50.7%70 / 83
53.5% best60 / 83
30.1% best50 / 77
24.0%62 / 77
87.1% best7 / 73
86.2%11 / 73
78.5% best23 / 58
75.8%33 / 58
85.0% best19 / 57
83.9%23 / 57
45.9%29 / 51
54.8% best17 / 51
77.5%42 / 50
78.4% best40 / 50