Kimi K2.5 vs Kimi K2 Thinking
Change modelsKimi K2.5Moonshot AI
Kimi K2 ThinkingMoonshot AI
IndexAll categories, out of 100.
51.6 ±2.2 best
47.8 ±8.6
GapIndex points below the top.
Top
−3.8 · no clear gap
CoverageIndex weight that has results.
100% best
65%
SpeedOutput tokens per second.
22/s
40/s best
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-01
2025-11
Input / 1MUS dollars per 1M input tokens.
$0.50 best
$0.60
Output / 1MUS dollars per 1M output tokens.
$2.50
$2.50
1K chats2K tokens in, 500 out, each.
$2.25 best
$2.45
1K documents20K tokens in, 1K out, each.
$12.50 best
$14.50
1K agent tasks200K tokens in, 20K out, each.
$150.00 best
$170.00
ContextMaximum tokens in one request.
262K
262K
AgenticMulti-step tasks with tools.
48.9 best
47.8
CodingCode writing and repair.
53.2 best
44.5
ReasoningLogic problems and puzzles.
51.1
—
MultimodalTasks with images and text.
51.6
—
KnowledgeFacts and expert knowledge.
53.2 best
48.9
MultilingualTasks in many languages.
52.1
—
InstructionTasks with strict prompt rules.
52.8
—
MathMath problems.
53.1 best
50.0
Benchmarks
5 sharedOnly Kimi K2.5: 84Only Kimi K2 Thinking: 17
Kimi K2.5
Kimi K2 Thinking