Inference
EN
Moonshot AI

Kimi K2.5 Thinking

Kimi K2.5 Thinking is a model from Moonshot AI. 11 benchmarks count toward its index, in 5 categories.

  • Input: text
RankedEnough results for an index.
IndexAll categories, out of 100.52.6±7.4Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.—Median 262K
EloLMArena rating and rank.1445±3 #66Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted48.3145 / 218
CodingCode writing and repair.5 counted51.6127 / 209
ReasoningLogic problems and puzzles.None counted——
MultimodalTasks with images and text.1 counted64.9—
KnowledgeFacts and expert knowledge.2 counted53.5109 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.1 counted57.3—

Results

11 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
AIMEMath95.6% ±0.38 / 95
MMMU ProMultimodal84.3% ±0.926 / 92
LiveCodeBenchCoding83.9% ±1.044 / 142
MMLU ProKnowledge85.9% ±0.352 / 135
GPQA Diamond · Vals AIKnowledge84.1% ±2.152 / 135
IOI v1Coding17.7% ±6.126 / 61
Terminal-Bench 2.0 · Vals AIAgentic40.4% ±5.232 / 66
SWE-benchCoding70.0% ±2.161 / 87
Vibe Code Bench v1.1Coding17.5% ±3.396 / 123
Code MigrationCoding7.0% ±1.367 / 77
Terminal-Bench 2.1 · Vals AIAgentic41.9% ±0.766 / 75
CorpFin v2Finance68.3% ±0.910 / 131
MedQAHealthcare94.4% ±0.217 / 95
TaxEval v2Finance74.2% ±0.840 / 143
MortgageTaxFinance66.5% ±0.934 / 97
CaseLaw v2Legal58.7% ±0.728 / 53
Finance Agent v1.1Finance50.6% ±2.828 / 51
MedCodeHealthcare39.3% ±2.169 / 106
MedScribeHealthcare76.4% ±2.076 / 109
Legal Research BenchLegal15.9% ±2.570 / 90
Finance Agent (v2)Finance35.8% ±2.068 / 79
EMBFinance28.5% ±2.265 / 75

Sources

BenchLMCC BY-NC 4.0Vals AINo licence stated

History

5 releases
Index53.7 → 52.6
23 Oct 202612 Nov 2026

Questions

How good is Kimi K2.5 Thinking?
Its index is 52.6 out of 100. That is place 123 of 283 ranked models.
What is Kimi K2.5 Thinking best at?
Knowledge: 53.5 out of 100, place 109 of 268.

Badge

Kimi K2.5 Thinking on Inference360: place 123 of 283, index 52.6
HTML
<a href="https://inference360.ai/models/moonshot-ai-kimi-k2-5-thinking"><img src="https://inference360.ai/badges/moonshot-ai-kimi-k2-5-thinking.svg" alt="Kimi K2.5 Thinking on Inference360: place 123 of 283, index 52.6" height="20"></a>
Markdown
[![Kimi K2.5 Thinking on Inference360: place 123 of 283, index 52.6](https://inference360.ai/badges/moonshot-ai-kimi-k2-5-thinking.svg)](https://inference360.ai/models/moonshot-ai-kimi-k2-5-thinking)

More from Moonshot AI

All models
Kimi K3.1 Pro75.7Kimi K366.9Kimi K2.658.2Kimi K2.7 Code57.7Kimi K2.5 (Reasoning)53.0