Kimi K2.6 vs Kimi K2.5

Change models
Kimi K2.6Moonshot AI
Kimi K2.5Moonshot AI
IndexAll categories, out of 100.
58.2 ±2.9 best
52.2 ±2.2
GapIndex points below the top.
Top
−6.0
CoverageIndex weight that has results.
95%
100% best
SpeedOutput tokens per second.
81/s best
36/s
EloLMArena rating and rank.
1455
—
ReleasedMonth of release.
2026-04
2026-01
Input / 1MUS dollars per 1M input tokens.
$0.65
$0.532 best
Output / 1MUS dollars per 1M output tokens.
$3.41
$3.32 best
1K chats2K tokens in, 500 out, each.
$3.01
$2.73 best
1K documents20K tokens in, 1K out, each.
$16.41
$13.97 best
1K agent tasks200K tokens in, 20K out, each.
$198.20
$172.90 best
ContextMaximum tokens in one request.
262K
262K
AgenticMulti-step tasks with tools.
59.5 best
49.4
CodingCode writing and repair.
59.3 best
53.9
ReasoningLogic problems and puzzles.
53.8 best
51.6
MultimodalTasks with images and text.
57.6 best
52.4
KnowledgeFacts and expert knowledge.
56.1 best
53.7
MultilingualTasks in many languages.
—
52.7
InstructionTasks with strict prompt rules.
60.0 best
53.5
MathMath problems.
60.0 best
53.8

Benchmarks

39 shared
Only Kimi K2.6: 37Only Kimi K2.5: 50
Kimi K2.6
Kimi K2.5
37.5% best57 / 222
30.7%74 / 222
81.0% best35 / 217
78.0%70 / 217
90.8% best37 / 217
87.6%61 / 217
8.0% best75 / 214
3.1%95 / 214
96.1% best37 / 200
92.2%54 / 200
91.1% best27 / 199
87.9%53 / 199
32.6%72 / 164
35.2% best65 / 164
95.9%13 / 149
95.9%13 / 149
76.0% best15 / 146
70.2%44 / 146
90.5% best26 / 141
87.6%41 / 141
26.9% best67 / 133
17.0%83 / 133
80.2% best19 / 114
76.8%38 / 114
79.4% best26 / 113
75.4%49 / 113
34.9% best58 / 91
34.3%59 / 91
90.5% best26 / 84
87.6%47 / 84
58.6% best38 / 83
50.7%70 / 83
34.7% best42 / 77
30.1%48 / 77
39.0% best9 / 69
27.9%21 / 69
BrowseCompAgentic
83.2% best26 / 63
60.6%53 / 63
89.6% best10 / 59
85.0%19 / 59