Skip to content
AI Atlas
ModelDeprecatedOpen weightsIdentity probable

Kimi K2 Thinking

Moonshot AIfamily · Kimihuggingface.co/moonshotai/Kimi-K2-Thinking

Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning. Built on the trillion-parameter Mixture-of-Experts (MoE) architecture introduced in...

Open in Graph
data quality72

Updated 23 min ago · first seen 11 Sept 2026

model_01M294WW95XPFV3AGY9BV6JMYW

Overview

Identity

Canonical model
Yesidentity confidence: mediumOne row per real model release. Artifacts (checkpoints, quantisations, conversions) and folded evaluation variants point here.
Official checkpoints
official_checkpoints = hf_repo identifiers carried by the model itself; artifacts are separate entities pointing here through canonical_id.
Artifacts
None recordedSeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
Provider deployments
2
API aliases
kimi-k2-thinkingmoonshotai/kimi-k2-thinkingIdentifiers under which providers and evaluators refer to this model.
Folded evaluation variants
0Effort / thinking variants (…-high, …-non-reasoning) are result configurations of this model, not separate models. Their old URLs redirect here.

Openness

Open weights— weights downloadable under Other; 7 dimensions unknown.

Weights downloadable under a permissive or Creative Commons licence allowing commercial use; code or data may be missing.

  • Weights

    Yes

  • Inference code

    —

  • Training code

    —

  • Training data

    —

  • Dataset

    —

  • Commercial use

    —

  • Redistribution

    —

  • Derivatives

    —

Licence: Other (unclassified licence) (unknown · stated as “other”)

dimensions marked null are unknown, not false

Key facts

Release date

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Status

Source:Artificial AnalysisT2observed 5 d agomedium

Openrouter id

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Architecture

Architecture

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Model type

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Parameters

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Library name

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Pipeline tag

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium

Hugging Face repo

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Capabilities

Modalities

Modalities
text
Input
text
Output
text

Capabilities

  • Tool calling

    Yes

    OpenRouter public model & pricing listing · T2

  • Structured output

    Yes

    OpenRouter public model & pricing listing · T2

  • Reasoning

    Yes

    OpenRouter public model & pricing listing · T2

  • Vision

    Unavailable

  • Audio

    Unavailable

  • Fine-tuning available

    Unavailable

Context window

Source:OpenRouter public model & pricing listingT2observed 5 d agomedium

Max output

Source:OpenRouter public model & pricing listingT2observed 4 d agomedium

Benchmarks13

Compare with another model →

Comparable same task and conditions · Partially comparable same task, conditions differ (effort, temperature, judge) · Not comparable different variant or metric

Benchmark results grouped by comparability group
Benchmark · groupBest scoreTrustConfigurationResultsvs leaderEvaluatedSource
Terminal-Benchagentic · accuracy · variant=hard · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvarianthardreasoningonconditions differ across rows → partially comparable2−34.8 ptvs gpt-5.6-solobs. 16 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · variant=Telecom · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantTelecomreasoningonconditions differ across rows → partially comparable1non-primary groupobs. 16 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysis1−6.14 ptvs Z.ai GLM 5.2obs. 11 Sept 2026artificialanalysis.aiT2
SWE-bench Verifiedcoding · resolved · board=Verified · system=mini-SWE-agentOfficial boardboardVerifiedsystemmini-SWE-agent1−13.4 ptvs Claude Opus 4.510 Dec 2025swebench.comT2
Artificial Analysis Intelligence Indexcomposite · indexIndependentreasoningonversion4.3conditions differ across rows → partially comparable2−31.3vs Claude Fable 5.1obs. 16 Sept 2026artificialanalysis.aiT2
IFBenchinstruction-following · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−15.2 ptvs Grok 4.3obs. 16 Sept 2026artificialanalysis.aiT2
Humanity's Last Examknowledge · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoningonconditions differ across rows → partially comparable2−35.3 ptvs Claude Fable 5.1obs. 16 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantDiamondreasoningonconditions differ across rows → partially comparable1non-primary groupobs. 16 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=GPQA Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantGPQA Diamond1−12.4 ptvs gpt-6-astraobs. 11 Sept 2026artificialanalysis.aiT2

Current rows only, grouped by benchmark → canonical metric → comparability group (task configuration). Effort variants folded into this model appear as rows of the same group. 13 current rows in total. “vs leader” compares with the current leader of the benchmark's primary group only; other groups are not directly comparable. Comparability rules →

Providers & Pricing2

All offers in the price terminal →
Provider deployments, cheapest output first
ProviderContextInput / 1MCached inOutput / 1MStatusObservedSource
Moonshot AI Platformcheapest outputmoonshotai/kimi-k2-thinking262.1Kout 100.3K$0.15active5 d agosince 11 Sept 2026openrouter.aiT2
OpenRoutermoonshotai/kimi-k2-thinking262.1Kout 100.3K$0.15active28 min agosince 12 Sept 2026openrouter.aiT2

USD per 1M tokens as published by each provider; native units (per-request fees, flex/priority tiers) are kept verbatim. Rows are append-only — every price change is kept in the history below. Cost of a workload →

Price history

Step lines per provider; amber markers are recorded changes. Click a marker or a row for the evidence behind that price.

Output price · USD / 1M tokens 2 providers

Output price history of Kimi K2 Thinking$0$1$2$3Sept 26Sept 26Sept 26Sept 26Sept 26Moonshot AI Platform: first observed → $2.5 · 11 Sept 2026OpenRouter: first observed → $2.5 · 12 Sept 2026
  • Moonshot AI Platform
  • OpenRouter
  • OpenRouterfirst observed $2.512 Sept 2026
  • Moonshot AI Platformfirst observed $2.511 Sept 2026

Input price · USD / 1M tokens 2 providers

Input price history of Kimi K2 Thinking$0$0.20$0.40$0.60$0.80Sept 26Sept 26Sept 26Sept 26Sept 26Moonshot AI Platform: first observed → $0.60 · 11 Sept 2026OpenRouter: first observed → $0.60 · 12 Sept 2026
  • Moonshot AI Platform
  • OpenRouter
  • OpenRouterfirst observed $0.6012 Sept 2026
  • Moonshot AI Platformfirst observed $0.6011 Sept 2026

Hardware fit37

Estimated

1 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
NVIDIA DGX B2004bit1,440 GB590.7 GB est.Yes
Apple M3 Ultra4bit—590.7 GB est.No
Mac Studio (Apple M5 Ultra)4bit—590.7 GB est.No
Apple M2 Ultra4bit—590.7 GB est.No
Apple M1 Ultra4bit—590.7 GB est.No
Apple M3 Max4bit—590.7 GB est.No
Apple M4 Max4bit—590.7 GB est.No
Mac Studio (Apple M5 Max)4bit—590.7 GB est.No
MacBook Pro (Apple M5 Max)4bit—590.7 GB est.No
Apple M2 Max4bit—590.7 GB est.No
Apple M1 Max4bit—590.7 GB est.No
Apple M4 Pro4bit—590.7 GB est.No
Mac mini (Apple M5 Pro)4bit—590.7 GB est.No
MacBook Pro (Apple M5 Pro)4bit—590.7 GB est.No
Apple M3 Pro4bit—590.7 GB est.No
Apple M1 Pro4bit—590.7 GB est.No
Apple M2 Pro4bit—590.7 GB est.No
Apple M44bit—590.7 GB est.No
iMac (Apple M4)4bit—590.7 GB est.No
Mac mini (Apple M6)4bit—590.7 GB est.No
MacBook Air (Apple M5)4bit—590.7 GB est.No
MacBook Pro (Apple M5)4bit—590.7 GB est.No
Apple M24bit—590.7 GB est.No
Apple M34bit—590.7 GB est.No
Apple M14bit—590.7 GB est.No
NVIDIA GeForce RTX 30904bit24 GB590.7 GB est.No
NVIDIA GeForce RTX 40904bit24 GB590.7 GB est.No
NVIDIA GeForce RTX 50904bit32 GB590.7 GB est.No
NVIDIA A100 80GB4bit80 GB590.7 GB est.No
NVIDIA H100 SXM4bit80 GB590.7 GB est.No
NVIDIA H100 NVL4bit94 GB590.7 GB est.No
NVIDIA DGX Spark4bit128 GB590.7 GB est.No
NVIDIA H2004bit141 GB590.7 GB est.No
NVIDIA H200 NVL4bit141 GB590.7 GB est.No
NVIDIA B2004bit180 GB590.7 GB est.No
AMD Instinct MI300X4bit192 GB590.7 GB est.No
AMD Instinct MI325X4bit256 GB590.7 GB est.No
Assumptions (7)
  • Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
  • bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
  • KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
  • A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
  • Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
  • Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
  • Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

Versions & Artifacts0

Version history

Max output3 changes

11 Sept 2026→12 Sept 2026→12 Sept 2026→13 Sept 2026current

Context window2 changes

11 Sept 2026→11 Sept 2026→12 Sept 2026current

Licensefirst observation only

11 Sept 2026current

Opennessfirst observation only

11 Sept 2026current

Parametersfirst observation only

11 Sept 2026current

Statusfirst observation only

11 Sept 2026current

Each hop is a claim: click a value for its source, tier and observation time. Nothing is overwritten — a new observation closes the previous claim.

Artifacts 0

No artifact (checkpoint, quantisation, conversion or packaging) points to this model yet.

Timeline14

Full timeline →
  • Max output changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: max output tokens changed from 100352 to 98304

    Max output100.3K tokens→98.3K tokensopenrouter
    openrouter.ai
  • Max output changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: max output tokens changed from 235929 to 100352

    Max output235.9K tokens→100.3K tokensopenrouter
    openrouter.ai
  • Price changedModelKimi K2 ThinkingMoonshot AI

    OpenRouter changed pricing for Kimi K2 Thinking: $0.6 in / $2.5 out per 1M tokens → $0.6 in / $2.5 out per 1M tokens

    $0.60 in / $2.5 out→$0.60 in / $2.5 outopenrouter
    openrouter.ai
  • Max output changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: max output tokens changed from 100352 to 235929

    Max output100.3K tokens→235.9K tokensopenrouter
    openrouter.ai
  • Listed by providerModelKimi K2 ThinkingMoonshot AI

    OpenRouter lists Kimi K2 Thinking at $0.6 in / $2.5 out per 1M tokens

    openrouter
    openrouter.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 31.06% on Terminal-Bench

    artificial_analysis
    artificialanalysis.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 92.98% on τ²-bench

    artificial_analysis
    artificialanalysis.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 68.1% on IFBench

    artificial_analysis
    artificialanalysis.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 23.82% on Humanity's Last Exam

    artificial_analysis
    artificialanalysis.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 83.84% on GPQA Diamond

    artificial_analysis
    artificialanalysis.ai
  • Benchmark resultModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking scores 22.02 on Artificial Analysis Intelligence Index

    artificial_analysis
    artificialanalysis.ai
  • Context window changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: context length changed from 256000 to 262144

    Context window256K tokens→262.1K tokensopenrouter
    openrouter.ai
  • Release date changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: release date changed from 2025-11-04 to 2025-11-06

    Release date4 Nov 2025→6 Nov 2025openrouter
    openrouter.ai
  • Release date changedModelKimi K2 ThinkingMoonshot AI

    Kimi K2 Thinking: release date changed from 2025-11-06 to 2025-11-04

    Release date6 Nov 2025→4 Nov 2025huggingface
    huggingface.co

Change history

Temporal, append-only claims: a new observation closes the previous claim instead of overwriting it. Rewind the record with the as-of picker.
11 claims · 1 propertiesShow all properties

Aa median output tokens per secondmetric.aa_median_output_tokens_per_second11

Claim history for Aa median output tokens per second
ValueValid from → toStatusSourceConfidenceExtractor
121.7 → currentcurrentArtificial AnalysisT2mediumdeterministic
122.1 → supersededArtificial AnalysisT2mediumdeterministic
122.9 → supersededArtificial AnalysisT2mediumdeterministic
123.2 → supersededArtificial AnalysisT2mediumdeterministic
123.6 → supersededArtificial AnalysisT2mediumdeterministic
123.4 → supersededArtificial AnalysisT2mediumdeterministic
123.7 → supersededArtificial AnalysisT2mediumdeterministic
122.8 → supersededArtificial AnalysisT2mediumdeterministic
121.7 → supersededArtificial AnalysisT2mediumdeterministic
119.4 → supersededArtificial AnalysisT2mediumdeterministic
115.2 → supersededArtificial AnalysisT2mediumdeterministic

Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →

Provenance

Attributed facts

45

Source tiers

T245

Freshest observation

23 min ago

Conflicts

None

Source documents 6

Source documents
SourceDocumentTypeTierLast observedSnapshots
Artificial Analysisartificialanalysis.ai/leaderboards/models leaderboardT2· Quality secondary23 min ago18
OpenRouter public model & pricing listingopenrouter.ai/api/v1/models catalogueT2· Quality secondary28 min ago76
Hugging Face Hub (public pages, model cards, papers)huggingface.co/moonshotai/Kimi-K2-Thinking/raw/main/README.md model_cardT2· Quality secondary39 min ago1
Hugging Face Hub (public pages, model cards, papers)huggingface.co/moonshotai/Kimi-K2-Thinking model_pageT2· Quality secondary1 h ago30
Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=moonshotai&p=0&sort=downloads listingT2· Quality secondary2 h ago32
SWE-bench leaderboardsswebench.com/ leaderboardT2· Quality secondary21 h ago1

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.

Data quality (72/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →