Skip to content
aiqbench

Moonshot · model passport

MoonshotAI: Kimi K3

kimi-k3
  • Open weights
  • Kimi K3 license licence
  • 2.78T params
  • 1,048,576 context
  • Released 2026-06-13(HF repo created)
12/27axes covered
Overall · LMArena
1,487#15 of 396
Best percentile
98.0Hard prompts
Cohort standing
77.3median 59.7

What the data says

Benchmarked on 12 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 98 pct — Hard prompts (1,516)
  • 97.7 pct — Coding (1,538)
  • 97.2 pct — Instruction following (1,488)

Uneven: 41.6 percentile points between its best and worst axis — 98 pct on Hard prompts versus 56.4 pct on Agent Arena.

  • Hard prompts98 pct1,516Verify →
  • Coding97.7 pct1,538Verify →
  • Instruction following97.2 pct1,488Verify →
  • Overall96.5 pct1,487Verify →
  • WebDev Arena92.1 pct1,655Verify →
  • GPQA Diamond91.9 pct93.1%Verify →
  • Factuality Arena86 pct1,471Verify →
  • AIME (math)84.5 pct97.2%Verify →
  • SimpleBench72.4 pct60.7%Verify →
  • SimpleQA Verified69.5 pct50.6%Verify →
  • ARC-AGI-267.4 pct60.4%Verify →
  • Agent Arena56.4 pct0.019Verify →
  • Vision ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • FrontierMathNot benchmarked
  • Humanity's Last ExamNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

Repositorymoonshotai/Kimi-K3
Parameters2.78T
ArchitectureKimiK3ForConditionalGeneration
LicenceKimi K3 license

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "moonshotai/Kimi-K3"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDmoonshotai/kimi-k3
Context length1,048,576 tokens
Max output943,718 tokens
Acceptstext, image, video
Input price$0.25 per 1M tokens
Cached input price$0.21 per 1M tokens
Output price$11.94 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin-pPresence penaltyReasoningReasoning effortRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-kTop logprobsTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"moonshotai/kimi-k3","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • Kimi K3 (Max) stands at 77.3 (raw mean 84.1) — measured on 12 of 27 axes.

    standingtemplateVerify →

  • Kimi K3 (Max) stands at 77.3 versus a cohort median of 59.7 across 206 models released within 180 days — above the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →