Kimi K3

Leaderboards

Where Kimi K3 sits among the top 20 in each category.

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Test results

Every test Kimi K3 has a published result on, out of 100%, with its rank among the models that took it.

• GPQA

  #8 / 256

  Graduate-level science questions written to be hard to look up.

  93.5%

  Self-reported

• Humanity's Last Exam

  #12 / 108

  Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions

  56.0%

  Self-reported · In the text score· In the vision score

• MMMU-Pro

  #8 / 73

  College-level questions that need both an image and text to answer.

  81.6%

  Self-reported · In the vision score

• BrowseComp

  #2 / 68

  Finding hard-to-locate facts by browsing the web.

  91.2%

  Self-reported · In the agents score

• CharXiv-R

  #2 / 60

  Reading and reasoning about charts from scientific papers.

  91.3%

  Self-reported

• DeepSWE 1.1

  #21 / 50

  DeepSWE 1.1 evaluates software engineering agents using the mini-swe-agent harness.

  69.0%

  Self-reported · In the coding score

• Terminal-Bench 2.1

  #6 / 46

  Terminal-Bench 2.1 is an updated release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal. It evaluates how well models handle real-world, end-to-end tasks autonomously, including compiling code, training models, setting up servers, system administration, data science workflows, and security tasks.

  88.3%

  Self-reported

• Toolathlon

  #7 / 42

  Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.

  73.2%

  Self-reported · In the agents score

• MCP Atlas

  #3 / 38

  MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across complex multi-step tasks.

  84.2%

  Self-reported

• MathVision

  #1 / 38

  MATH-Vision is a dataset designed to measure multimodal mathematical reasoning capabilities. It focuses on evaluating how well models can solve mathematical problems that require both visual understanding and mathematical reasoning, bridging the gap between visual and mathematical domains.

  97.8%

  Self-reported

• AutomationBench

  #14 / 24

  AutomationBench is a tool-use benchmark that evaluates AI agents on automating real-world workflows, testing their ability to orchestrate tools and complete multi-step automation tasks.

  30.8%

  Self-reported · In the agents score

• Program Bench

  #4 / 18

  Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation. It spans 200 tasks from small CLI tools to large systems such as FFmpeg and SQLite, with submissions judged against more than 248,000 fuzz-generated behavioral tests.

  77.8%

  Self-reported · In the coding score

• BabyVision

  #4 / 16

  A benchmark for early-stage visual reasoning and perception on child-like vision tasks.

  85.7%

  Self-reported

• FrontierSWE

  #2 / 16

  FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.

  81.2%

  Self-reported

• ZEROBench

  #5 / 15

  ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.

  0.4%

  Self-reported

• DeepSWE

  #3 / 13

  DeepSWE is a software engineering agent benchmark evaluated with the mini-swe-agent harness, where each task is solved in an isolated container with no internet access. It measures an agent's ability to autonomously resolve real-world coding issues end to end.

  67.5%

  Self-reported

• OfficeQA Pro

  #7 / 12

  OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.

  63.3%

  Self-reported

• APEX-Agents

  #3 / 11

  APEX-Agents is a benchmark evaluating AI agents on long horizon professional tasks that require sustained reasoning, planning, and execution across complex multi-step workflows.

  37.6%

  Self-reported

• DeepSearchQA

  #1 / 10

  DeepSearchQA is a benchmark for evaluating deep search and question-answering capabilities, testing models' ability to perform multi-hop reasoning and information retrieval across complex knowledge domains.

  95.0%

  Self-reported

• SWE-Marathon

  #7 / 10

  SWE-Marathon is an ultra-long-horizon software engineering benchmark covering tasks such as building compilers, optimizing kernels, and developing production-grade services. It measures whether agents can sustain quality across extremely long engineering trajectories.

  42.0%

  Self-reported

• Job Bench

  #8 / 9

  Job Bench evaluates AI agents on realistic professional tasks that require multi-step planning, research, and production of work artifacts.

  52.9%

  Self-reported

• PostTrainBench

  #5 / 9

  PostTrainBench evaluates a model's ability to autonomously post-train base models. Given pretrain-only base models, the agent must complete the full pipeline of data synthesis, training, evaluation, and iteration within a time budget, scored across downstream benchmarks such as AIME2025, BFCL, GPQA Main, GSM8K, and HumanEval.

  36.6%

  Self-reported

• WorldVQA

  #3 / 7

  WorldVQA is a benchmark designed to evaluate atomic vision-centric world knowledge. It assesses models' ability to understand and reason about visual elements representing real-world knowledge.

  51.0%

  Self-reported

• AA-Briefcase

  #2 / 4

  AA-Briefcase is an Artificial Analysis evaluation of AI systems on professional knowledge-work tasks, reported as an Elo score.

  51.6%

  Self-reported

• MMMU-Pro (with tools)

  #2 / 4

  MMMU-Pro variant evaluated with tool access enabled.

  83.4%

  Self-reported

• MLS-Bench Lite

  #1 / 3

  MLS-Bench Lite is the official 30-task subset of MLS-Bench for evaluating whether AI systems can invent generalizable and scalable machine learning methods across LLM pretraining and post-training, robotics, world models, computer vision, reinforcement learning, optimization, ML systems, and AI for Science.

  48.3%

  Self-reported

• Kimi Code Bench v2

  #1 / 2

  Kimi Code Bench v2 is Moonshot AI's in-house benchmark for evaluating coding agents on realistic software engineering tasks across 10+ mainstream programming languages and a production tech stack spanning backend services, infrastructure, performance engineering, systems programming, security, frontend development, and ML/data engineering.

  72.9%

  Unverified

• OmniDocBench

  #1 / 2

  OmniDocBench evaluates multimodal models on document understanding tasks such as OCR, layout parsing, and structured document comprehension.

  91.1%

  Unverified

• PerceptionBench

  #2 / 2

  PerceptionBench is Moonshot AI's internal benchmark for evaluating atomic visual perception capabilities.

  58.5%

  Unverified

• SpreadsheetBench 2

  #1 / 2

  SpreadsheetBench 2 evaluates office automation agents on spreadsheet analysis, reasoning, and manipulation tasks.

  34.8%

  Unverified

• DECK-Bench

  #1 / 1

  DECK-Bench is Moonshot AI's internal evaluation of agents on creating and reasoning over presentation-style knowledge-work artifacts.

  73.5%

  Unverified