Skip to content

Moonshot AI · Chat

Kimi K2.6

Moonshot AI 🇨🇳ChatKimi familyOpen weights. One of its sources says the weights are publicly available.The maker currently plans to deprecate it within the next 180 days (16 Apr 2027). Plans can change.

Newer version: Kimi K3.1. Our catalogue has no release date for Kimi K3.1. We compared the version names in our catalogue.

From Empiriolabs's listing:
Kimi K2.6 is a Moonshot multimodal reasoning model with 256K context, strong coding, and text, image, and video inputs.
llminfo IQ
132.6
Max context
262K
Listings
162

Evidence and sources

International API storefront is Singapore-based; principal operating entity 北京月之暗面科技有限公司 (Beijing).

🤗 moonshotai/Kimi-K2.6 Model architecture, as its sources state it llminfo IQ 132.6, ranked among chat models. From 305 benchmarks. Plenty of data behind it.How llminfo IQ is computedCoding IQ 112.9, 95% interval 106–119. From 23 benchmark families across 11 sources. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.How Coding IQ is computedFair value (50.4). Costs per task about what its benchmark results predict. From 55 benchmarks with measured task costs. Plenty of data behind it.How Value for money worksYap score: tokens used per coding task vs. models with the same resultsTypical (48 ±14). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it. By effort: default 48 · none 56.Median 23,752 output tokens per task at default effort.How the Yap score worksReleased 14 Apr 2026Maker HQ: CN (source)
All rankings: #52 on NanoGPT recently #90 on OpenRouter today #94 on OpenRouter this month #97 on OpenRouter this weekSources: Writer Pricing, EmpirioLabs data, LiteLLM data, Models.dev data, Morpheus Inference Marketplace data, Hugging FacePrimary data from Writer Pricing from EmpirioLabs: primary mode, description from LiteLLM: deprecation date from Models.dev: weights availability from Morpheus Inference Marketplace: architecture from Hugging Face: release date
Tags from its sources:instructtoolsreasoningvisiondocumentopen-sourcemoelong-contextcoding

Advantages and trade-offs

Compared with 10 chat models, picked mainly by popularity and IQ

Advantages

Open weights available

The weights are public. 3 of the 9 peers that say either way have open weights.

Trade-offs

Deprecation scheduled

The maker plans to deprecate it on 16 Apr 2027, within the next 180 days.

We only make a claim when there's enough data on models of the same type. Updated 22 Nov, 04:34 UTC.

Input & output

as stated by the model's providers
Text
Images
Video
Files
Text

Capabilities

Supported operations and provider-stated features

VisionReasoningFunction callingStructured outputPrompt cachingWeb search

Listings

 Moonshot AIkimi-k2.6Sources: GenAI Prices data, LiteLLM data, Models.dev data262K$0.95cache read $0.16$4.00——
 Jzstokenkimi-k2.6Source: Models.dev data262K$0.1557cache read $0.0156$0.7784cache write $0.1557——
 Routing Runkimi-k2.6Source: Models.dev data200K$0.275$1.10——
 OpenRouterInference by Baidumoonshotai/kimi-k2.6 · fp4Source: OpenRouter data docs262K$0.3895cache read $0.0656$1.6481922ms
 NanoGPTTEE/kimi-k2.6 · teeSources: Models.dev data, NanoGPT262K$0.39cache read $0.039$1.80——
 Chutesmoonshotai/Kimi-K2.6-TEE · tee · int4Sources: Chutes data, Models.dev data262K$0.39cache read $0.039$1.80——
 Cortecs AIkimi-k2.6Sources: Cortecs EU Router data, Models.dev data256K——
 Hugging FaceInference by Inceptronmoonshotai/Kimi-K2.6Source: Hugging Face Inference Providers data262K$0.5648$2.788——
 Inceptronmoonshotai/Kimi-K2.6 · int4Sources: Inceptron data, Models.dev data262K$0.5648cache read $0.19$2.788——
 Kenarikimi-k2-6Source: Models.dev data262K$0.5714cache read $0.2857$2.8571——

12 listings without a price are hidden.

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Human preference

Arena ELO60 related benchmarks · 60 results
Source: LMArena data · Nov 2026
BenchmarkScore
Arena ELO
Chinese
Creative Writing
Document
Show 56 more benchmarksHide 56 benchmarks
BenchmarkScore
Document Style-Controlled
English
Exclude Ties
Expert
French
German
Hard Prompts
Hard Prompts English
Industry Business and Management and Financial Operations
Industry Entertainment and Sports and Media
Industry Life and Physical and Social Science
Industry Mathematical
Industry Medicine and Healthcare
Industry Software and It Services
Industry Writing and Literature and Language
Japanese
Korean
Longer Query
Multi Turn
Non English
Polish
Russian
Spanish
Style-Controlled
Text Factuality
Text Factuality - Chinese
Text Factuality - Creative Writing
Text Factuality - English
Text Factuality - Exclude Ties
Text Factuality - Expert
Text Factuality - French
Text Factuality - Hard Prompts
Text Factuality - Hard Prompts English
Text Factuality - Industry Business and Management and Financial Operations
Text Factuality - Industry Entertainment and Sports and Media
Text Factuality - Industry Life and Physical and Social Science
Text Factuality - Industry Mathematical
Text Factuality - Industry Medicine and Healthcare
Text Factuality - Industry Software and It Services
Text Factuality - Industry Writing and Literature and Language
Text Factuality - Japanese
Text Factuality - Longer Query
Text Factuality - Multi Turn
Text Factuality - Non English
Text Factuality - Russian
Text Factuality - Spanish
Vision
Vision - Chinese
Vision - Creative Writing Vision
Vision - Diagram
Vision - English
Vision - Homework
Vision - Humor
Vision Style-Controlled
Design Arena EloSource: Design Arena Leaderboards data · Nov 2026

Agentic

Apex Agents2 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · 452-task subset
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
BrowseComp3 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · Agent Swarm
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · context management · discard-all strategy
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
Design Arena EloSource: Design Arena Leaderboards data · Nov 2026
MCP AtlasKimi K2.7 Code official comparison table · Kimi K2.6 · Kimi Code CLI · thinking · temp 1.0 · top_p .95 · 262,144 context · 100 tool-call budget · 32K max tokens/step · avg@3 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
MCP-MarkSource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
MCPMark VerifiedSource: MCPMark Verified leaderboard docs
OpenHandsSource: OpenHands Index data
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
SWE-Bench Pro2 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · in-house SWE-agent-derived harness · avg@10
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
Source: OpenRouter benchmark exports
20 Apr 2026 version · 90-day results
Across 18 providers lowest 62.3% median 72.8% highest 74%OpenRouter's automatic routing 73%
Show results from 18 providers and OpenRouter's automatic routingHide results from 18 providers and OpenRouter's automatic routing
  • AtlasCloud
  • Cloudflare
  • CoreWeave
  • Moonshot AI
  • Inceptron
  • OpenRouter's automatic routing
  • Crusoe
  • DigitalOcean
  • Novita
  • Phala
  • Baidu
  • Parasail
  • SiliconFlow
  • DeepInfra
  • Venice
  • Chutes
  • StreamLake
  • GMICloud
  • Decart
Terminal-Bench 2.0 · Terminus-2Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · Terminus-2 · JSON parser · preserve thinking · avg@10 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Coding

Arena ELOSource: LMArena data · Nov 2026
Claw-EvalSource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Kimi Code Bench 2.0Kimi K2.7 Code official comparison table · Kimi K2.6 · Kimi Code CLI · thinking · temp 1.0 · top_p .95 · 262,144 context · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · v6 · avg@10
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
MLS-Bench-LiteKimi K2.7 Code official comparison table · Kimi K2.6 · Kimi Code CLI · thinking · temp 1.0 · top_p .95 · 262,144 context · official 30-task subset · five-hour agent budget · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
SciCode2 results
Source: Epoch AI Benchmarking Hub data
Setting not stated
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · avg@10
SWE-bench Verified5 related benchmarks · 7 results
SWE-bench Verified
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · in-house SWE-agent-derived harness · avg@10
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
SWE-bench Multilingual
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · in-house SWE-agent-derived harness · avg@10
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
ValsSource: Vals AI

Math

HMMT Feb 263 related benchmarks · 3 results
HMMT Feb 26
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
HMMT 2026 Feb
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context
Hmmt Feb 2026
Source: Hugging Face Official Benchmarks data · Reported by the model's developer. The source relaying it did not verify it. claim
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
MathArenaSource: MathArena data
MathVision3 related benchmarks · 3 results
MathVision
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
With CI
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · with Python · max 50 steps · avg@3
Without CI
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no Python · avg@3

Science

GPQASource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
GPQA Diamond21 results
Source: Epoch AI Benchmarking Hub data
Setting not stated
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context
Kimi K2.6 official launch blog · headline comparison table
Source: OpenRouter benchmark exports
20 Apr 2026 version · 90-day results
Across 17 providers lowest 73.5% median 89.3% highest 90.5%OpenRouter's automatic routing 88.3%
Show results from 17 providers and OpenRouter's automatic routingHide results from 17 providers and OpenRouter's automatic routing
  • Crusoe
  • Inceptron
  • Cloudflare
  • Phala
  • Novita
  • SiliconFlow
  • Moonshot AI
  • Decart
  • Baidu
  • DigitalOcean
  • Venice
  • Parasail
  • CoreWeave
  • OpenRouter's automatic routing
  • Chutes
  • StreamLake
  • AtlasCloud
  • DeepInfra

General

Reasoning

DeepSearchQA2 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no context management; over-length tasks counted failed
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · full set · no tools · max generation 98,304
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · text-only subset · no tools · footnote result
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · full set · search + code interpreter + browser · 262,144 max generation · context management
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · text-only subset · with tools · footnote result
OJBenchSource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Toolathlon VerifiedKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
V*Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim

Long context

Vision

CharXiv-R3 related benchmarks · 3 results
CharXiv-R
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
CharXiv (RQ) · With CI
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · RQ · with Python · max 50 steps · avg@3
CharXiv (RQ) · Without CI
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · RQ · no Python · avg@3
MMMU-Pro3 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no Python · official protocol · avg@3
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · with Python · max 50 steps · avg@3
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Factuality

Safety

PhareSource: Phare

Instruction following

Multilingual

OCR

PerceptionBench · OCR accuracybenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge · Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Knowledge

Speech

Other

AIME262 results
Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context
Kimi K2.6 official launch blog · headline comparison table
BabyVision · With CIKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · with Python · max 50 steps · avg@3 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
BabyVision · Without CIKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no Python · avg@3 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Claw-Eval v1.18 related benchmarks · 8 results
Overall · Pass³
Reported by Claw-Eval, Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Claw-Eval v1.1 · overall · 300 tasks · N=3 · grader: Gemini 3 Flash
General (161) · Pass@3
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · general · 161 tasks · N=3 · grader: Gemini 3 Flash
General (161) · Pass³
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · general · 161 tasks · N=3 · grader: Gemini 3 Flash
Multi-turn (38) · Pass@3
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · multi-turn · 38 tasks · N=3 · grader + user-agent: Claude Opus 4.6
Multi-turn (38) · Pass³
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · multi-turn · 38 tasks · N=3 · grader + user-agent: Claude Opus 4.6
Multimodal (101) · Pass@3
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · multimodal · 101 tasks · N=3 · grader: Gemini 3 Flash
Show 2 more resultsHide 2 results
Multimodal (101) · Pass³
Reported by Claw-Eval · via BenchBoard docs
Claw-Eval v1.1 · multimodal · 101 tasks · N=3 · grader: Gemini 3 Flash
Overall · Pass@3
Reported by Claw-Eval, Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Claw-Eval v1.1 · overall · 300 tasks · N=3 · grader: Gemini 3 Flash
DeepSearchQA · AccuracyKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no context management; over-length tasks counted failed · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
E-Commerce Bench7 related benchmarks · 7 results
Mean final assets
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
AnchorRatio · repeat-order pricing
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
BadSpend · fraudulent-supplier spend
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
Controllable return
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
CSE+ · honest-deal negotiation
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
Drawdown / peak assets
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
Show 1 more resultHide 1 result
Yuan per tool call
Reported by E-Commerce Bench authors / Alibaba Qwen · via BenchBoard docs
benchmark-owner 365-day merchant simulation · five independent episodes · up to four stores · initial balance ¥100,000 · thinking enabled where supported; effort not specified · exact API snapshot not stated for unversioned labels
KimiReported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
MCPMark-VerifiedKimi K2.7 Code official comparison table · Kimi K2.6 · Kimi Code CLI · thinking · temp 1.0 · top_p .95 · 262,144 context · 100-step tool-call budget · 32K max tokens/step · avg@3 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
NovaPPTBenchSenseNova 6.7 Flash-Lite official benchmark chart · internal benchmark · provider states it will be open-sourced · Reported by SenseTime / OpenSenseNova · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
OJBench · PythonKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · Python · avg@10 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
PerceptionBench9 related benchmarks · 9 results
Overall accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Attribute accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Comparison accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Contextual integration accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Counting accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Depth and 3D perception accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Show 3 more resultsHide 3 results
Fine-grained recognition accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Localization accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
Visual relation accuracy
Reported by Moonshot AI / PerceptionBench authors · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge
ProgramBench · Metric UnspecifiedKimi K2.7 Code official comparison table · Kimi K2.6 · Kimi Code CLI · thinking · temp 1.0 · top_p .95 · 262,144 context · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
SkillsBench3 related benchmarks · 3 results
1.1 · With Skills
Reported by SkillsBench / BenchFlow · via BenchBoard docs
OpenHands · 87 tasks · 3 trials/task · max reasoning effort · with Skills
1.1 · Without Skills
Reported by SkillsBench / BenchFlow · via BenchBoard docs
OpenHands · 87 tasks · 3 trials/task · max reasoning effort · without Skills
Self-contained subset
Reported by Qwen · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Qwen3.7 release comparison table · OpenCode · 78 tasks excluding 9 external-API tasks · five-run average · table effort=Thinking
V* · With CIKimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · with Python · max 50 steps · avg@3 · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Vals12 related benchmarks · 14 results
EMB
Source: Vals AI
CaseLaw v2
Source: Vals AI
CorpFin v2
Source: Vals AI
LegalBench
Source: Vals AI
MedCode
Source: Clinical Benchmarks data docs · Sep 2026
Setting not stated
Source: Vals AI
Setting not stated
Show 7 more resultsHide 7 results
MedScribe
Source: Clinical Benchmarks data docs · Sep 2026
Setting not stated
Source: Vals AI
Setting not stated
MortgageTax
Source: Vals AI
WideSearch item F1Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · hide-tool-result context management · Reported by Moonshot AI · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Data credits: Hugging Face official benchmark leaderboards; licence varies by benchmark · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · Data by LLM Stats · LMArena, CC BY 4.0 (licence) · Bangla LLM Bench, Md Deluair Hossen. Source tasks: Hendrycks et al. 2021 (MMLU), OpenAI MMMLU, Bandarkar et al. 2024 (Belebele), Shi et al. 2023 (MGSM), Nahin et al. 2025 (TituLLMs). Public numeric community results; results licence not separately stated. · BullshitBench by Peter Gostev, MIT (licence) · Clinical Benchmarks (clinicalbenchmarks.ai), CC BY 4.0; fractions converted to percentages; source documents retain their own terms (licence) · EQ-Bench (Samuel J. Paech) — cite paech2023eqbench · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence) · MathArena (ETH SRI / INSAIT), MIT — cite dekoninck2026matharena (licence) · MCPMark (eval-sys), mcpmark.ai Verified leaderboard. Single-run Pass@1 fractions shown as percentages. · MineBench, MIT (licence) · OpenHands Index (All Hands AI), Apache-2.0 (licence) · Phare (Giskard), CC BY 4.0 (licence) · Vectara Hallucination Leaderboard, Apache-2.0 (licence)

Also known as: @cf/moonshotai/kimi-k2.6, aihubmix/kimi-k2.6, azure_ai/fw-kimi-k2.6, azure_ai/kimi-k2.6, baseten/moonshotai/kimi-k2.6, cline-pass/kimi-k2.6, cloudflare/@cf/moonshotai/kimi-k2.6, deepinfra/moonshotai/kimi-k2.6, e2ee-kimi-k2.6, empiriolabs/kimi-k2-6, fireworks_ai/accounts/fireworks/models/kimi-k2p6, fireworks_ai/accounts/fireworks/routers/kimi-k2p6-fast, fireworks_ai/kimi-k2p6, fireworks_ai/kimi-k2p6-fast, gmi/moonshotai/kimi-k2.6, inceptron/kimi-k2.6, kimi k2.6, kimi-k2-6, kimi-k2.6-fast, kimi-k2.6-nitro, kimi-k2.6:web, kimi-k2p6, kimi/kimi-k2.6, moonshot/kimi-k2-6, moonshot/kimi-k2.6, moonshotai/kimi-k2.6, moonshotai/kimi-k2.6-tee, moonshotai/kimi-k2.6:thinking, nebius/moonshotai/kimi-k2.6, novita/kimi-k2.6, novita/moonshotai/kimi-k2.6, nvidia_nim/moonshotai/kimi-k2.6, parasail-kimi-k26, pro/moonshotai/kimi-k2.6, replicate/moonshotai/kimi-k2.6, streamlake/kimi-k2.6, tee/kimi-k2.6, tencent/kimi-k2.6, tensormesh/moonshotai/kimi-k2.6, together_ai/moonshotai/kimi-k2.6, wandb/moonshotai/kimi-k2.6

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works