Examenos

Kimi K3 Benchmarks, Specifications & Availability

Explore Kimi K3 from Moonshot AI: published specifications, source-linked vendor benchmarks, independent evaluator coverage and recorded pricing when available.

Compare Kimi K3 with other models →Explore data coverage

Published specifications

Provider
Moonshot AI
Access
Open weights
License
Kimi K3 License
Context window
1M
Total parameters
2.8T
Active parameters
104B
Released
2026-07-16
Modalities
text, image, video
Family
Kimi K3

Model card · Announcement · Website · OpenRouter

Model notes

2.8T MoE / 104B active; native multimodal (text/image/video per OR; GH summary also lists text+image). Ignore OR :batch twin for main pricing. Tech report arXiv:2607.24653. Open weights verified on Hugging Face https://huggingface.co/moonshotai/Kimi-K3 (2026-09-25).

Pricing · OpenRouter

Dated cached OpenRouter rates in USD per 1M tokens. Open the dashboard for live enhancements. Per-metric endpoint minima can refer to different providers; they are not a guaranteed combined rate from one endpoint.

Recorded pricing tiers
TierInput / 1MOutput / 1MCached input / 1MCache write / 1MDate & source
Default[object Object][object Object][object Object]—2026-10-03 · OpenRouter source
Recorded pricing notes

OR prompt/completion/input_cache_read ×1e6; ignore :batch twin. No -contribute sibling.; min-healthy endpoint minima 2026-10-01; discount flag 0.15 on Phala (undocumented, not applied); min-healthy endpoint minima 2026-10-01; min-healthy endpoint minima 2026-10-02; discount flag 0.25 on Phala (undocumented, not applied); min-healthy endpoint minima 2026-10-03; discount flag 0.25 on Phala (undocumented, not applied)

Official / vendor benchmarks

Default headline records. Own-vendor, peer-vendor and third-party provenance remain visible in evidence; configurations may differ.

Official / vendor headline scores; expand evidence for every record
Benchmark / evaluatorHeadline scoreEvidence
GPQA Diamond93.5%max effort
All 1 recorded result & sources

93.5% · raw 93.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

GPQA Diamond; reasoning_effort=max, temp=1.0, top_p=0.95

https://github.com/MoonshotAI/Kimi-K3
CritPt23.4%max effort
All 1 recorded result & sources

23.4% · raw 23.4 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

CritPt; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
AA-LCR74.7%max effort
All 1 recorded result & sources

74.7% · raw 74.7 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

AA-LCR; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
Humanity's Last Exam43.5%max effort
All 1 recorded result & sources

43.5% · raw 43.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

HLE-Full without tools (with tools 56.0 stored separately)

https://github.com/MoonshotAI/Kimi-K3
Humanity's Last Exam (w/ tools)56%max effort
All 1 recorded result & sources

56% · raw 56 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

HLE-Full with general tools

https://github.com/MoonshotAI/Kimi-K3
DeepSWE v1.167.5%max effort
All 2 recorded results & sources

67.5% · raw 67.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

DeepSWE v1.1 tasks; Kimi Code harness (mini-SWE-agent board reports 67.3)

https://github.com/MoonshotAI/Kimi-K3

67.5% · raw 67.5 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

https://naive.ai/en/research/
ProgramBench77.8%max effort
All 1 recorded result & sources

77.8% · raw 77.8 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

ProgramBench; Kimi Code harness

https://github.com/MoonshotAI/Kimi-K3
Terminal-Bench 2.188.3%max effort
All 2 recorded results & sources

88.3% · raw 88.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

Terminal-Bench 2.1; Kimi Code harness

https://github.com/MoonshotAI/Kimi-K3

88.3% · raw 88.3 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

https://naive.ai/en/research/
FrontierSWE81.2%max effort
All 2 recorded results & sources

81.2% · raw 81.2 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

FrontierSWE dominance; Kimi Code harness; as of 2026-07-16

https://github.com/MoonshotAI/Kimi-K3

81.2% · raw 81.2 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

https://naive.ai/en/research/
SWE-Marathon42%max effort
All 1 recorded result & sources

42% · raw 42 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

SWE-Marathon; Claude Code harness; H20-calibrated branch pre-v1.1

https://github.com/MoonshotAI/Kimi-K3
PostTrainBench36.6%max effort
All 2 recorded results & sources

36.6% · raw 36.6 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

PostTrainBench; Harbor + Claude Code harness; avg of 3 on H20

https://github.com/MoonshotAI/Kimi-K3

36.6% · raw 36.6 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

https://naive.ai/en/research/
MLS-Bench-Lite48.3%max effort
All 1 recorded result & sources

48.3% · raw 48.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MLS-Bench-Lite; Kimi Code harness

https://github.com/MoonshotAI/Kimi-K3
SciCode58.7%max effort
All 1 recorded result & sources

58.7% · raw 58.7 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

SciCode; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
Kimi Code Bench 2.072.9%max effort
All 1 recorded result & sources

72.9% · raw 72.9 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

Kimi Code Bench 2.0 (in-house); Kimi Code harness (73.7 with Claude Code noted)

https://github.com/MoonshotAI/Kimi-K3
BrowseComp91.2%max effort
All 1 recorded result & sources

91.2% · raw 91.2 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

BrowseComp with context-compaction at 300K (90.4 with full 1M / no compaction)

https://github.com/MoonshotAI/Kimi-K3
DeepSearchQA95%max effort
All 1 recorded result & sources

95% · raw 95 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

DeepSearchQA F1

https://github.com/MoonshotAI/Kimi-K3
ResearchRubrics76.2%max effort
All 1 recorded result & sources

76.2% · raw 76.2 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

ResearchRubrics

https://github.com/MoonshotAI/Kimi-K3
GDPval-AA v21686max effort
All 1 recorded result & sources

1686 · raw 1686 Elo

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

GDPval-AA v2 Elo; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
Toolathlon-Verified76.5%max effort
All 1 recorded result & sources

76.5% · raw 76.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

Toolathlon-Verified

https://github.com/MoonshotAI/Kimi-K3
MCPMark-Verified94.5%max effort
All 1 recorded result & sources

94.5% · raw 94.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MCPMark-Verified

https://github.com/MoonshotAI/Kimi-K3
MCP Atlas84.2%max effort
All 1 recorded result & sources

84.2% · raw 84.2 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MCP-Atlas 500-task public subset, 100-turn limit, Gemini 3.1 Pro judge

https://github.com/MoonshotAI/Kimi-K3
AutomationBench v1.0.630.8%max effort
All 1 recorded result & sources

30.8% · raw 30.8 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

AutomationBench 600-task public subset (mapped to automationbench-v1.0.6)

https://github.com/MoonshotAI/Kimi-K3
JobBench54.3%max effort
All 1 recorded result & sources

54.3% · raw 54.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

JobBench

https://github.com/MoonshotAI/Kimi-K3
AA Briefcase v1.11548max effort
All 1 recorded result & sources

1548 · raw 1548 Elo

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

AA-Briefcase Elo; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
Agents' Last Exam28.3%max effort
All 2 recorded results & sources

28.3% · raw 28.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

Agents' Last Exam primary pass-rate; Kimi Code harness; from official leaderboard as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3

28.3% · raw 28.3 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI.

https://naive.ai/en/research/
APEX-Agents41%max effort
All 1 recorded result & sources

41% · raw 41 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

APEX-Agents; Moonshot cites APEX leaderboard as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
OfficeQA Pro63.3%max effort
All 1 recorded result & sources

63.3% · raw 63.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

OfficeQA Pro; entire PDF corpus as images; Claude Code harness

https://github.com/MoonshotAI/Kimi-K3
SpreadsheetBench 234.8%max effort
All 1 recorded result & sources

34.8% · raw 34.8 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

SpreadsheetBench 2; Claude Code harness

https://github.com/MoonshotAI/Kimi-K3
OSWorld-Verified84.8%max effort
All 1 recorded result & sources

84.8% · raw 84.8 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

OSWorld-Verified

https://github.com/MoonshotAI/Kimi-K3
OSWorld 2.058.3%max effort
All 1 recorded result & sources

58.3% · raw 58.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

OSWorld 2.0

https://github.com/MoonshotAI/Kimi-K3
SaaS-Bench60.1%max effort
All 1 recorded result & sources

60.1% · raw 60.1 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

SaaS-Bench

https://github.com/MoonshotAI/Kimi-K3
τ³-Bench Banking33.4%max effort
All 1 recorded result & sources

33.4% · raw 33.4 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

τ³-Banking; Moonshot cites Artificial Analysis as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
Harvey Legal Agent94.6%max effort
All 1 recorded result & sources

94.6% · raw 94.6 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

Harvey Lab-AA criterion pass rate; Moonshot cites AA as of 2026-07-23

https://github.com/MoonshotAI/Kimi-K3
CorpFin v271.6%max effort
All 1 recorded result & sources

71.6% · raw 71.6 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

CorpFin v2; Moonshot cites Vals AI

https://github.com/MoonshotAI/Kimi-K3
Vals Finance Agent v254.4%max effort
All 1 recorded result & sources

54.4% · raw 54.4 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

Finance Agent v2; Moonshot cites Vals AI

https://github.com/MoonshotAI/Kimi-K3
Legal Research Bench44.2%max effort
All 1 recorded result & sources

44.2% · raw 44.2 %

Headline · max effort · Peer vendor

Source/record date: 2026-07-16

Legal Research Bench; Moonshot cites Vals AI

https://github.com/MoonshotAI/Kimi-K3
WorldVQA51%max effort
All 1 recorded result & sources

51% · raw 51 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

WorldVQA ForceAnswer

https://github.com/MoonshotAI/Kimi-K3
OmniDocBench 1.591.1%max effort
All 1 recorded result & sources

91.1% · raw 91.1 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

OmniDocBench (mapped to omnidocbench-1.5; version not further specified on GH)

https://github.com/MoonshotAI/Kimi-K3
PerceptionBench58.5%max effort
All 1 recorded result & sources

58.5% · raw 58.5 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

PerceptionBench (in-house atomic visual perception)

https://github.com/MoonshotAI/Kimi-K3
Video-MME90%max effort
All 1 recorded result & sources

90% · raw 90 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

Video-MME (w. subtitles)

https://github.com/MoonshotAI/Kimi-K3
MMVU82.1%max effort
All 1 recorded result & sources

82.1% · raw 82.1 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MMVU

https://github.com/MoonshotAI/Kimi-K3
BabyVision (w/ tools)85.7%max effort
All 1 recorded result & sources

85.7% · raw 85.7 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

BabyVision w/ python

https://github.com/MoonshotAI/Kimi-K3
MMMU Pro (no tools)81.6%max effort
All 1 recorded result & sources

81.6% · raw 81.6 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MMMU-Pro without tools (with tools 83.4 stored separately)

https://github.com/MoonshotAI/Kimi-K3
MMMU Pro (with tools)83.4%max effort
All 1 recorded result & sources

83.4% · raw 83.4 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MMMU-Pro with Python tools

https://github.com/MoonshotAI/Kimi-K3
CharXiv RQ91.3%max effort
All 2 recorded results & sources

91.3% · raw 91.3 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

CharXiv (RQ) with tools (without tools 84.8 noted)

https://github.com/MoonshotAI/Kimi-K3

84.8% · raw 84.8 %

Alternative · max effort · Own vendor

Source/record date: 2026-07-16

CharXiv (RQ) without tools

https://github.com/MoonshotAI/Kimi-K3
MathVision97.8%max effort
All 2 recorded results & sources

97.8% · raw 97.8 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

MathVision with tools (without tools 94.3 noted)

https://github.com/MoonshotAI/Kimi-K3

94.3% · raw 94.3 %

Alternative · max effort · Own vendor

Source/record date: 2026-07-16

MathVision without tools

https://github.com/MoonshotAI/Kimi-K3
ZeroBench-main (w/ tools)41%max effort
All 1 recorded result & sources

41% · raw 41 %

Headline · max effort · Own vendor

Source/record date: 2026-07-16

ZeroBench pass@5 with tools (without tools 23.0 noted)

https://github.com/MoonshotAI/Kimi-K3
Gray Swan IPI52.7%unknown effort
All 1 recorded result & sources

52.7% · raw 52.7 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-30

As reported by Google Gemini 4 Argon Gray Swan chart (K=15 attack success rate)

https://storage.googleapis.com/gweb-uniblog-publish-prod/images/gemini_4_cyber_evals_gray_swan_i.width-1200.format-webp.webp
CyberGym80%unknown effort
All 1 recorded result & sources

80% · raw 80 %

Headline · unknown effort · Third-party

Source/record date: 2026-09-30

Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed.

https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym
SWE-Pro64.8%unknown effort
All 1 recorded result & sources

64.8% · raw 64.84 %

Headline · unknown effort · Third-party

Source/record date: 2026-09-30

Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed.

https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym

Independent evaluators

Evaluator harnesses are distinct from vendor measurements. Missing coverage is not a failed test.

Independent evaluator headline scores; expand evidence for every record
Benchmark / evaluatorHeadline scoreEvidence
Intelligence Index · Artificial Analysis44max effort
All 2 recorded results & sources

44 · raw 44 index

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

https://artificialanalysis.ai/models/kimi-k3

30 · raw 30 index

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

https://artificialanalysis.ai/models/kimi-k3-low
Output speed · Artificial Analysis34 tok/smax effort
All 2 recorded results & sources

34 tok/s · raw 34 tok/s

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

https://artificialanalysis.ai/models/kimi-k3

36 tok/s · raw 36 tok/s

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens.

https://artificialanalysis.ai/models/kimi-k3-low
Vals Index · Vals AI50.3%max effort
All 1 recorded result & sources

50.3% · raw 50.3 %

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Refreshed from current Vals leaderboard. Cost/test $6.38.

https://www.vals.ai/benchmarks/vals_index
Bugs fixed /105 · Bug Hunt Bench21 fixesunknown effort
All 1 recorded result & sources

21 fixes · raw 21 fixes

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Claude Code / OpenRouter; effort default; 1 runs; evaluation 2026-07-26. Best documented score for this effort in Oct 1 README. Headline: best documented model run.

https://github.com/phuryn/bug-hunt-bench
ARC-AGI-1 · ARC Prize94.5%max effort
All 3 recorded results & sources

94.5% · raw 94.5 %

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties.

https://arcprize.org/results/moonshot-kimi-k3

86.7% · raw 86.7 %

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

https://arcprize.org/results/moonshot-kimi-k3

65.7% · raw 65.7 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

https://arcprize.org/results/moonshot-kimi-k3
ARC-AGI-2 · ARC Prize60.4%max effort
All 3 recorded results & sources

60.4% · raw 60.4 %

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties.

https://arcprize.org/results/moonshot-kimi-k3

55% · raw 55 %

Alternative · high effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

https://arcprize.org/results/moonshot-kimi-k3

12.4% · raw 12.4 %

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Semi-Private verified matrix. Published verified configuration.

https://arcprize.org/results/moonshot-kimi-k3
Hard board % of roofline · KernelBench (community board)18.9%unknown effort
All 1 recorded result & sources

18.9% · raw 18.9 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-09-23

Kimi K3 (1M) kinetic-claude; Hard 6/6; Mega 9.79× 1/1; CUDA 10.6% 4/4

https://kernelbench.com/models/kinetic-0715-1m
CUDA board % of roofline · KernelBench (community board)10.6%unknown effort
All 1 recorded result & sources

10.6% · raw 10.6 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-09-23

Kimi K3 (1M); CUDA 4/4; Hard 18.9% already ingested

https://kernelbench.com/models/kinetic-0715-1m
Money gain · Andon Labs$4665.04unknown effort
All 2 recorded results & sources

$4665.04 · raw 4665.04 $

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Moonshot provider. Direct model-vendor deployment is headline among provider variants. VB2 final balance minus $500. Corrects prior plain Kimi 2693.75 gain, which came from the vb1 block rather than vb2.

https://andonlabs.com/evals/vending-bench-2

$4406.96 · raw 4406.96 $

Alternative · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Harness: Fireworks provider. VB2 final balance minus $500; alternate deployment.

https://andonlabs.com/evals/vending-bench-2
Blueprint Bench · Andon Labs29.5%unknown effort
All 1 recorded result & sources

29.5% · raw 29.5 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Blueprint-Bench 2 connectivity similarity; published fractional score multiplied by 100.

https://andonlabs.com/evals/blueprint-bench-2
Average Score · WeirdML v37.4%high effort
All 1 recorded result & sources

7.4% · raw 7.42 %

Headline · high effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant Kimi K3 (high, Fireworks); harness kimi_code 2.0.0; values from prepared data JSON; raw 0.074209; official 80/20 aggregate (area 500k-50M tokens + final best)

https://htihle.github.io/weirdml.html
Final Best Score · WeirdML v315.6%high effort
All 1 recorded result & sources

15.6% · raw 15.62 %

Headline · high effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant Kimi K3 (high, Fireworks); harness kimi_code 2.0.0; values from prepared data JSON; raw 0.156217; mean final best effective score

https://htihle.github.io/weirdml.html
Cost / Run · WeirdML v3$19.53high effort
All 1 recorded result & sources

$19.53 · raw 19.53 USD

Headline · high effort · Independent evaluator

Source/record date: 2026-10-02

WeirdML variant Kimi K3 (high, Fireworks); harness kimi_code 2.0.0; values from prepared data JSON; mean API cost per run, same task weighting as scores

https://htihle.github.io/weirdml.html
Vibe Code Bench v1.1 · Vals AI85%unknown effort
All 1 recorded result & sources

85% · raw 84.97 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-03

Refreshed from current Vals leaderboard. Harness: OpenHands. Cost/test $10.01.

https://www.vals.ai/benchmarks/vibe-code
Cost per Intelligence Index task · Artificial Analysis$2.00max effort
All 2 recorded results & sources

$2.00 · raw 2 USD

Headline · max effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

https://artificialanalysis.ai/models/kimi-k3

$1.15 · raw 1.15 USD

Alternative · low effort · Independent evaluator

Source/record date: 2026-10-03

Current AA model leaderboard;

https://artificialanalysis.ai/models/kimi-k3-low

Read how we select and source scores or the comparison guide.