# Kimi K3

Moonshot AI · Open Weight · rank 36 · bench-align-v5

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/models/kimi-3  
JSON: https://modelscale.dev/api/model/kimi-3

## Facts

| Field | Value | Evidence |
| --- | --- | --- |
| Canonical id | `kimi-3` | — |
| Overall score | 65.5 | Observed 2026-11-21 · source benchlm:models |
| Verified score (BenchLM): a separate BenchLM figure, not the overall score | 72 | Observed 2026-11-21 · source benchlm:models |
| Context window | 1.05M tokens | Observed 2026-11-21 · source benchlm:models |
| Max output (models.dev) | 1,048,576 tokens | Observed 2026-11-21 · source modelsdev:catalog |
| Modalities (models.dev) | Input: text, image, video · Output: text | Observed 2026-11-21 · source modelsdev:catalog |
| Release date | 2026-07-16 | Observed 2026-11-21 · source benchlm:models |
| Access type | Open Weight | — |
| Blended $/1M (75% input / 25% output) | $6.00 | Derived from the input and output rates below |
| Cost per successful task (LiveBench) | $0.35 | Observed 2026-11-21 · source livebench:cost |

## Capability evidence

Seven axes from the ranking source. An axis the source did not score is unavailable, not zero. Beside each score is BenchLM's separate verified score for that axis; the two are published side by side and are not interchangeable.

| Axis | Score | Evidence | Verified (BenchLM) | Verified evidence |
| --- | --- | --- | --- | --- |
| Agentic | 81.6 | Observed 2026-11-21 · source benchlm:models | 67.9 | Observed 2026-11-21 · source benchlm:models |
| Coding | 77.5 | Observed 2026-11-21 · source benchlm:models | 67.6 | Observed 2026-11-21 · source benchlm:models |
| Knowledge | 81.2 | Observed 2026-11-21 · source benchlm:models | 80 | Observed 2026-11-21 · source benchlm:models |
| Reasoning | 66.7 | Observed 2026-11-21 · source benchlm:models | 66.4 | Observed 2026-11-21 · source benchlm:models |
| Multimodal & Grounded | 85.8 | Observed 2026-11-21 · source benchlm:models | 85.8 | Observed 2026-11-21 · source benchlm:models |
| Instruction Following | Unavailable | Unavailable · source benchlm:models | Unavailable | Unavailable · source benchlm:models |
| Math | Unavailable | Unavailable · source benchlm:models | Unavailable | Unavailable · source benchlm:models |

## Runtime service evidence

Measured values with the date they were observed. Nothing is inferred from a sibling model or a provider claim. Regional or per-endpoint measurements appear only when the API supplies them; none are modelled.

| Measurement | Value | Observed | Last good | Evidence |
| --- | --- | --- | --- | --- |
| Time to first token | 56.25 s | 2026-11-21 | 2026-11-21 | Observed 2026-11-21 · source benchlm:speed |
| Throughput | 38 tok/s | 2026-11-21 | 2026-11-21 | Observed 2026-11-21 · source benchlm:speed |

## Endpoint and price matrix

Every published price component, including cache reads and writes.

| Component | USD | Evidence |
| --- | --- | --- |
| Input / 1M tokens | $3.00 | Observed 2026-11-21 · source benchlm:pricing |
| Output / 1M tokens | $15.00 | Observed 2026-11-21 · source benchlm:pricing |
| Cache read / 1M tokens | $0.30 | Observed 2026-11-21 · source benchlm:pricing |
| Cache write / 1M tokens | $3.00 | Observed 2026-11-21 · source modelsdev:catalog |
| Blended / 1M (75% input / 25% output) | $6.00 | Derived — from the input and output rates above; it has no source record of its own |
| Cost per successful task (LiveBench) | $0.35 | Observed 2026-11-21 · source livebench:cost |
| Batch input / 1M tokens (OpenRouter batch) | $2.28 | Observed 2026-11-21 · source openrouter:pricing |
| Batch output / 1M tokens (OpenRouter batch) | $11.40 | Observed 2026-11-21 · source openrouter:pricing |
| Batch cache read / 1M tokens (OpenRouter batch) | $0.23 | Observed 2026-11-21 · source openrouter:pricing |

Batch API rate (asynchronous jobs), from OpenRouter's batch listing.

**Pricing note.** Moonshot AI's official Kimi K3 pricing page lists $0.30 cache-hit input, $3.00 cache-miss input, and $15.00 output per million tokens for model ID kimi-k3, with an exact 1,048,576-token context window. OpenRouter independently lists route moonshotai/kimi-k3 with one Moonshot AI INT4 endpoint at the same rates and context length. Prices exclude applicable taxes. Moonshot published the full weights on Hugging Face under the Kimi K3 License (model card checked October 29, 2026).

## Workload-aware monthly cost example

10 conversations per day × 8 messages × 22 active days, 1200 input and 400 output tokens per message, no cache. Derived here from the published rates above by this site's own calculator — not a figure any source published.

| Field | Value |
| --- | --- |
| Modelled monthly cost | $16.90 |
| Modelled tokens | 2.82M |

## Benchmark record

107 matched benchmark rows with their published value, unit, and provenance.

### Knowledge

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| GPQA (Graduate-Level Google-Proof Q&A) | 93.5 | 448 questions | Graduate level | [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://arxiv.org/abs/2311.12022) |
| GPQA-D (GPQA Diamond) | 93.5 | Graduate-level science questions | Graduate level | [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://arxiv.org/abs/2311.12022) |
| HLE (Humanity's Last Exam) | 56 | Expert-level questions | Frontier expert level | [Humanity's Last Exam](https://lastexam.ai/) |
| Artificial Analysis Intelligence Index | 43.6 | Cross-benchmark intelligence index | Display-only external reference | [Artificial Analysis](https://artificialanalysis.ai/) |
| AA-GPQA Diamond (Artificial Analysis GPQA Diamond) | 93.5 | Graduate-level science questions | Graduate-level science reasoning | [Artificial Analysis GPQA Diamond Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/gpqa-diamond) |
| AA-HLE (Artificial Analysis Humanity's Last Exam) | 46.9 | Expert-level questions | Frontier expert reasoning | [Artificial Analysis Humanity's Last Exam Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/hle) |
| AA-Omniscience Index (Artificial Analysis Omniscience Index) | 19.7 | Knowledge questions | Broad factual knowledge | [AA-Omniscience: Knowledge and Hallucination Benchmark](https://artificialanalysis.ai/evaluations/omniscience) |
| AA-Omniscience Accuracy (Artificial Analysis Omniscience Accuracy) | 47.6 | Knowledge questions | Broad knowledge | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/grok-4-3) |
| AA-Omniscience Hallucination Rate (Artificial Analysis Omniscience Hallucination Rate) | 53.2 | Knowledge questions | Factuality | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/grok-4-3) |
| HLE w/o tools (Humanity's Last Exam without tools) | 43.5 | Expert-level questions | Frontier expert level | [Introducing GPT-5.4 mini and nano](https://openai.com/index/introducing-gpt-5-4-mini-and-nano/) |
| GPQA Diamond (Vals) (GPQA Diamond, Vals AI run) | 92.9 | Graduate-level science questions | Expert reasoning | [Vals AI GPQA Diamond, Vals AI run leaderboard](https://www.vals.ai/benchmarks/gpqa) |
| MMLU-Pro (Vals) (MMLU-Pro, Vals AI run) | 88.0 | Academic multiple-choice questions | Broad academic knowledge | [Vals AI MMLU-Pro, Vals AI run leaderboard](https://www.vals.ai/benchmarks/mmlu_pro) |

### Coding

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| VulcanBench v3 | 73.7 | 23 post-cutoff repository tasks in the v3 report | Professional multi-file software engineering | [VulcanBench](https://github.com/morganlinton/VulcanBench/tree/main) |
| OpenHarmony Bench v1.0 | 57.3 | 153 app-development and bug-fix tasks | End-to-end OpenHarmony application development | [OpenHarmony Bench: Evaluating LLMs and Coding Agents on OpenHarmony App Development](https://arxiv.org/abs/2608.16022) |
| ProgramBench (ProgramBench: Can Language Models Rebuild Programs From Scratch?) | 77.8 | 200 program reconstruction tasks | Full-repository software architecture | [ProgramBench: Can Language Models Rebuild Programs From Scratch?](https://programbench.com/static/paper.pdf) |
| PostTrainBench (lab runs) (PostTrainBench lab-run results with no stated version) | 36.6 | Post-training base models under the PostTrainBench budget | Autonomous ML engineering | [PostTrainBench](https://posttrainbench.com/) |
| FrontierSWE | 81.2 | 17 ultra-long-horizon engineering and research tasks | Ultra-long-horizon frontier software engineering | [FrontierSWE: Benchmarking coding agents at the limits of human abilities](https://www.frontierswe.com/blog) |
| FrontierSWE v2 | 25.9 | 34 ultra-long-horizon engineering and research tasks | Ultra-long-horizon frontier software engineering | [FrontierSWE v2](https://www.frontierswe.com/blog/v2) |
| Bug Hunt Bench | 21.0 | 105 planted bugs across two production TypeScript repositories | Blind production-repository bug finding and repair | [Bug Hunt Bench method, caveats, and definitions](https://bughunt.productcompass.pm/method) |
| Kimi Code Bench v2 | 72.9 | Realistic coding-agent tasks | Production software engineering | [Kimi K2.7 Code](https://huggingface.co/moonshotai/Kimi-K2.7-Code) |
| MLS-Bench Lite | 48.3 | 30 machine-learning research tasks | ML research and systems engineering | [MLS-Bench](https://mls-bench.com/) |
| AA Coding Index (Artificial Analysis Coding Index) | 76.2 | Cross-benchmark coding index | Display-only external reference | [Artificial Analysis model leaderboards](https://artificialanalysis.ai/leaderboards/models) |
| AA-SciCode (Artificial Analysis SciCode) | 59.5 | Scientific coding subproblems | Scientific programming | [Artificial Analysis SciCode Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/scicode) |
| LiveCodeBench (Vals) (LiveCodeBench, Vals AI run) | 87.2 | Competitive programming problems (easy, medium, hard) | Frontier coding | [Vals AI LiveCodeBench, Vals AI run leaderboard](https://www.vals.ai/benchmarks/lcb) |
| SWE-bench (Vals) (SWE-bench, Vals AI run) | 93.4 | Real repository issues by human time bucket | Frontier coding agents | [Vals AI SWE-bench, Vals AI run leaderboard](https://www.vals.ai/benchmarks/swebench) |
| DeepSWE | 67.5 | 113 software engineering tasks across 91 repositories and 5 languages | Long-horizon software engineering | [DeepSWE benchmark blog](https://deepswe.datacurve.ai/blog) |
| PostTrainBench v1.1 | 32.0 | Post-training Qwen3 1.7B and 4B, SmolLM3 3B, and Gemma3 4B | Frontier agent evaluation | [PostTrainBench v1.1 leaderboard and methodology](https://posttrainbench.com/?version=v1.1) |
| FrontierCode 2.0 | 18.9 | 132 private software-engineering tasks | Frontier coding-agent quality | [FrontierCode 2.0](https://cognition.com/blog/frontier-code-2.0) |
| ScarfBench (ScarfBench: Cross-Framework Application Migration in Enterprise Java) | 61.3 | 204 framework migration tasks across 34 enterprise Java applications | Whole-application framework migration | [ScarfBench: A Benchmark for Cross-Framework Application Migration in Enterprise Java](https://arxiv.org/abs/2605.06754) |
| PostTrainBench (PostTrainBench v1.2) | 31.5 | Post-training four base models on six benchmarks | Autonomous ML engineering | [PostTrainBench](https://posttrainbench.com/) |

### Reasoning

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| ARC-AGI-1 (ARC-AGI-1 Semi-Private Evaluation) | 94.50 | Semi-private ARC-AGI-1 evaluation set | Abstract visual reasoning | [ARC Prize leaderboard](https://arcprize.org/) |
| ARC-AGI-2 (Abstraction and Reasoning Corpus for AGI v2) | 60.416667 | Visual pattern completion and abstract reasoning | Expert-level — hardest public reasoning benchmark | [ARC-AGI-2: A Harder General Intelligence Benchmark](https://arcprize.org/arc-agi/2/) |
| AA-LCR (Artificial Analysis Long Context Reasoning) | 88.7 | Long-context reasoning tasks | Long-context reasoning | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/grok-4-3) |
| MLCR-AA (Medical Long Context Reasoning (MLCR-AA)) | 38.3 | Long, fragmented medical-record reasoning | Long-context medical reasoning | [Medical Long Context Reasoning (MLCR-AA)](https://artificialanalysis.ai/evaluations/mlcr-aa) |
| CritPt (Critical Physics Tasks) | 23.4 | Research-level physics questions | Research-level physics reasoning | [CritPt Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/critpt) |

### Agentic

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| AA Briefcase (Artificial Analysis Briefcase) | 1499 | Professional knowledge-work tasks | Professional work | [Artificial Analysis Briefcase Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/aa-briefcase) |
| AA AutomationBench (Artificial Analysis AutomationBench) | 58.3 | Business-process automation tasks | Agentic automation | [Artificial Analysis AutomationBench Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/automationbench-aa) |
| AA EnterpriseOps-Gym (Artificial Analysis EnterpriseOps-Gym) | 45.3 | Enterprise operations workflows | Enterprise agent operations | [Artificial Analysis EnterpriseOps-Gym Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/enterprise-ops-gym-aa) |
| AA Harvey LAB v1.0 (Artificial Analysis Harvey LAB-AA v1.0) | 94.6 | Legal agent tasks | Professional legal work | [Artificial Analysis Harvey LAB-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/harvey-lab-aa) |
| AA ITBench (Artificial Analysis ITBench-AA) | 47.7 | IT incident-response tasks | Enterprise IT operations | [Artificial Analysis ITBench-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/itbench-aa) |
| AA Tau3 Banking (Artificial Analysis Tau3-Banking) | 46.0 | Banking tool-use workflows | Agentic banking workflows | [Artificial Analysis Tau3-Banking Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/tau3-banking) |
| Terminal-Bench 2.0 | 88.3 | Terminal-based software tasks | Professional software engineering | [Terminal-Bench 2.0](https://www.tbench.ai/) |
| Terminal-Bench 2.1 | 88.3 | Terminal-based software-agent tasks | Professional software engineering | [Terminal-Bench](https://www.tbench.ai/) |
| BrowseComp | 91.2 | Research questions requiring browsing | Hard web research | [BrowseComp](https://openai.com/index/browsecomp/) |
| GDPval-AA | 1584 | Agentic real-world work tasks | Professional agentic workflows | [DeepSeek-V4 Technical Report](https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro/blob/main/DeepSeek_V4.pdf) |
| GDPval-AA (GDPval-AA normalized) | 51.6 | Economically valuable tasks | Professional agentic workflows | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/grok-4-3) |
| AA Agentic Index (Artificial Analysis Agentic Index) | 50.6 | Cross-benchmark agentic index | Display-only external reference | [Artificial Analysis model leaderboards](https://artificialanalysis.ai/leaderboards/models) |
| AA Terminal-Bench 4.0 (Artificial Analysis Terminal-Bench v4.0) | 12.6 | Terminal-based agent tasks | Agentic software engineering | [Artificial Analysis Terminal-Bench v4.0 Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/terminalbench-v4-0) |
| GDP.pdf (Artificial Analysis GDP.pdf) | 22.0 | Professional document-production tasks | Professional knowledge work | [GDP.pdf Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/gdp-pdf) |
| AA-AnalystAgent (Artificial Analysis AnalystAgent) | 38.8 | Spreadsheet and document analysis questions | Business and data analysis | [AA-AnalystAgent Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/aa-analyst-agent) |
| APEX-Agents-AA | 41.3 | 452 professional-services agent tasks | Long-horizon workplace agent tasks | [APEX-Agents-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/apex-agents-aa) |
| JobBench | 52.9 | 130 tasks across 35 occupations | Professional multi-source workflows | [JobBench: Aligning Agent Work With Human Will](https://arxiv.org/abs/2605.26329) |
| MCP Atlas | 84.2 | Tool-integrated agent tasks | Advanced tool use | [Introducing GPT-5.4 mini and nano](https://openai.com/index/introducing-gpt-5-4-mini-and-nano/) |
| Toolathlon-Verified | 73.2 | Verified multi-tool workflows | Advanced tool use | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| AutomationBench | 30.8 | 600 public automation tasks | Long-horizon automation | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| APEX-Agents | 37.6 | Professional-services agent tasks | Long-horizon professional work | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| SpreadsheetBench 2 | 34.8 | Spreadsheet analysis and editing tasks | Professional spreadsheet work | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| DECK-Bench (DECK-Bench (Internal)) | 73.5 | Internal presentation workflows | Professional presentation creation | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| DeepSearchQA | 95.0 | Agentic browsing and list-answer questions | Agentic web research | [Muse Spark Eval Methodology](https://ai.meta.com/static-resource/muse-spark-eval-methodology) |
| Terminal-Bench 2.1 (Vals) (Terminal-Bench 2.1, Vals AI run) | 80.9 | Difficult terminal tasks | Frontier agentic | [Vals AI Terminal-Bench 2.1, Vals AI run leaderboard](https://www.vals.ai/benchmarks/terminal-bench-2-1) |
| ApprenticeBench (ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job) | 18 | 100 vendor bills processed in sequence inside a simulated construction company | Long-horizon computer use with offline and online continual learning | [ApprenticeBench: a step change in AI's job readiness](https://neocognition.io/blog/apprentice-bench/) |
| Hermes Index (Hermes Index: agentic capability and cost per task inside Hermes Agent) | 39.2 | Four suites: Hermes Bench (150 tasks), TerminalBench 4 without its 4 GPU tasks, TerminalBench Science and SkillsBench | Frontier agentic | [Hermes Index: an agentic model leaderboard](https://portal.nousresearch.com/bench) |
| Hermes Bench (Hermes Bench: 150 agent tasks graded on the files and state the agent leaves behind) | 71.5 | 150 tasks across 25 categories | Frontier agentic | [Hermes Bench (Hermes Index)](https://portal.nousresearch.com/bench) |
| MazeBench (MazeBench: visual spatial reasoning in a 3D open world) | 1 | 100 hidden gems across 256 rooms in one open world | Long-horizon spatial planning, with single puzzles that can take over one hundred moves | [MazeBench Results](https://mazebench.com/blog?post=maze-bench-results) |
| Terminal-Bench 4.1 | 23.00 | 66 professional computer-work tasks | Frontier autonomous knowledge work | [Terminal-Bench 4.1](https://www.tbench.ai/news/terminal-bench-4-1) |
| AA Cyber Index (Artificial Analysis Cyber Index) | 41.3 | Three defensive cybersecurity evaluations | Real-world vulnerability discovery and patching | [Artificial Analysis Cyber Index](https://artificialanalysis.ai/evaluations/artificial-analysis-cyber-index) |
| CWE-Bench-AA (Artificial Analysis CWE-Bench-AA) | 53.3 | Real-world vulnerability patching tasks | Automated vulnerability remediation | [CWE-Bench-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/cwe-bench-aa) |
| DeepsecBench-AA (Artificial Analysis DeepsecBench-AA) | 12.5 | Vulnerability discovery in real codebases | Security code review | [DeepsecBench-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/deepsecbench-aa) |
| CyberGym-E2E-AA (Artificial Analysis CyberGym-E2E-AA) | 58.0 | 131 end-to-end vulnerability reproduction and patching tasks | Real-world cybersecurity | [CyberGym-E2E-AA Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/cybergym-e2e-aa) |
| Finch (Finch (FinWorkBench): Finance & Accounting across Spreadsheet-Centric Enterprise Workflows) | 77.3 | Spreadsheet-centric finance and accounting workflows | Enterprise finance work | [Finch: Benchmarking Finance & Accounting across Spreadsheet-Centric Enterprise Workflows](https://arxiv.org/abs/2512.13168) |
| b3 Agent Security (b3: Backbone Breaker Benchmark) | 88.0 | Threat snapshots of attacks on agent backends | Agent security | [The Backbone Breaker Benchmark](https://www.lakera.ai/blog/the-backbone-breaker-benchmark) |
| Cyber Misuse Refusal (Refusal of harmful cyber requests (JailbreakBench, StrongREJECT, AgentHarm)) | 94.0 | Cyber-related harmful prompts from three safety benchmarks | Cyber misuse safety | [Mistral Large 4 is now open weight](https://mistral.ai/news/mistral-large-4-oss/) |

### Multimodal & Grounded

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| MMMU-Pro (Massive Multi-discipline Multimodal Understanding Pro) | 81.6 | Multimodal academic reasoning | Frontier multimodal | [MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark](https://arxiv.org/abs/2409.02813) |
| AA-MMMU-Pro (Artificial Analysis MMMU-Pro) | 80.5 | Multimodal academic reasoning | Frontier multimodal | [Artificial Analysis MMMU-Pro Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/mmmu-pro) |
| Design Arena Website (Design Arena Website Elo) | 1335 | Website generation comparisons | Design and website generation | [OpenRouter Grok 4.3 benchmarks](https://openrouter.ai/x-ai/grok-4.3/benchmarks) |
| OfficeQA Pro | 63.3 | Document and spreadsheet tasks | Enterprise grounded reasoning | [OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning](https://arxiv.org/abs/2603.08655) |
| MathVision w/ Python (MathVision with Python) | 97.8 | Visual mathematics problems with Python | Advanced multimodal mathematics | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| BabyVision w/ Python (BabyVision with Python) | 85.7 | Visual perception tasks with Python | Fine-grained visual perception | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| ZeroBench w/ Python (ZeroBench_main with Python) | 41.0 | Visual reasoning questions with Python | Tool-augmented visual reasoning | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| WorldVQA ForceAnswer | 51.0 | Atomic visual world-knowledge questions | Fine-grained visual knowledge | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| OmniDocBench | 91.1 | Complex document-understanding tasks | Grounded document reasoning | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| PerceptionBench (PerceptionBench (Internal)) | 58.5 | Internal atomic visual-perception tasks | Fine-grained visual perception | [Kimi K3: Open Frontier Intelligence](https://www.kimi.com/blog/kimi-k3) |
| MMMU-Pro w/ Python (MMMU-Pro with Python) | 83.4 | Multimodal academic reasoning | Frontier multimodal | [Introducing GPT-5.4 mini and nano](https://openai.com/index/introducing-gpt-5-4-mini-and-nano/) |
| MathVision | 94.3 | Visually grounded math problems | Advanced multimodal mathematics | [Qwen3.6 launch benchmarks](https://qwen.ai/blog?id=qwen3.6) |
| ZeroBench | 23.0 | 100 visual reasoning questions | Tool-augmented visual reasoning | [Muse Spark Eval Methodology](https://ai.meta.com/static-resource/muse-spark-eval-methodology) |
| CharXiv (CharXiv Reasoning) | 91.3 | Scientific chart reasoning | Scientific visualization reasoning | [CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs](https://charxiv.github.io/) |
| CharXiv w/o tools (CharXiv Reasoning without tools) | 84.8 | Scientific chart reasoning (tool-free) | Scientific visualization reasoning | [CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs](https://charxiv.github.io/) |
| ExtractBench (LlamaIndex) (ExtractBench: schema-guided enterprise document extraction (LlamaIndex leaderboard)) | 83.17 | 370 documents, 4,869 pages, 8 business domains, 67 document types, each with a JSON Schema | Short documents are near solved; long record lists and grounding separate systems | [ExtractBench: A Benchmark for Schema-Guided Enterprise Document Extraction](https://arxiv.org/abs/2607.29677) |

### external

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| ExploitBench (ExploitBench v8-bench) | 32 | V8 exploit synthesis runs | Browser exploitation and cybersecurity | [ExploitBench](https://exploitbench.ai/) |
| ACE solved (ACE Cyber Range Challenges Solved) | 0 | 41 advanced cyber-range challenges | Advanced cyber operations | [UK AISI and CAISI Preliminary Assessment of Kimi K3's Cyber Capabilities](https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-preliminary-assessment-kimi-k3s-cyber-capabilities) |
| The Last Ones steps (The Last Ones Average Progress) | 17 | 32-step long-horizon cyber range | Long-horizon cyber operations | [UK AISI and CAISI Preliminary Assessment of Kimi K3's Cyber Capabilities](https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-preliminary-assessment-kimi-k3s-cyber-capabilities) |
| The Last Ones completion (The Last Ones Cyber Range Completion Rate) | 10 | 10 long-horizon runs | Long-horizon cyber operations | [UK AISI and CAISI Preliminary Assessment of Kimi K3's Cyber Capabilities](https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-preliminary-assessment-kimi-k3s-cyber-capabilities) |
| Chess Puzzles (Chess Puzzles (Epoch AI)) | 36.0 | 100 programmatically generated chess positions with a single best move | Novel chess positions, single best move | [Chess Puzzles (Epoch AI)](https://epoch.ai/benchmarks/chess-puzzles) |
| Mystery Game Puzzles (Mystery Game Puzzles (Epoch AI)) | 26.0 | 100 single-best-move positions from an undisclosed game | Unfamiliar game positions, single best move | [Mystery Game Puzzles (Epoch AI)](https://epoch.ai/benchmarks/mystery-game-puzzles) |
| Furniture Assembly (Furniture Assembly Benchmark (Epoch AI)) | 34.2 | 60 photos of three IKEA builds, some with deliberate assembly mistakes | Spatial reasoning from technical diagrams to real photos | [Furniture Assembly Benchmark (Epoch AI)](https://epoch.ai/benchmarks/furniture-assembly) |
| MirrorCode (MirrorCode (Epoch AI and METR)) | 54.4 | 30 reimplementation tasks (15 programs in two languages), three attempts each | Days-long software engineering without source code | [MirrorCode (Epoch AI)](https://epoch.ai/benchmarks/mirrorcode) |
| Epoch Automation Reports | 52.4 | 11 open-ended work tasks in five categories from Epoch AI's own work | Open-ended professional research and design work | [Epoch Automation Reports](https://epoch.ai/benchmarks/epoch-automation-reports) |
| Code2Video | 1481 | Creative briefs rendered as motion-graphics video | Code-driven motion design | [BenchLM](https://benchlm.ai/benchmarks/code2video) |
| Landing Page Arena | 1564 | 12 client briefs for small-business landing pages, rated by human votes | Marketing page design and copy | [BenchLM](https://benchlm.ai/benchmarks/landing-page-arena) |
| Pelican Arena | 1496 | One prompt: an animated SVG of a pelican riding a bicycle, rated by human votes | Vector drawing and animation in code | [BenchLM](https://benchlm.ai/benchmarks/pelican-arena) |
| ASCII Arena | 1301 | ASCII drawings of subjects from Google's Quick, Draw! dataset, rated by human votes | Text-character drawing | [BenchLM](https://benchlm.ai/benchmarks/ascii-arena) |
| BabaIsBench (BabaIsBench (Baba Is You intro levels, pass@1)) | 96 | 8 intro levels of Baba Is You, three attempts per level (Stage 0); 15 further levels run once (Stage 1) | Rule-rewriting puzzles that need planning over many moves | [Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost? (Quesma)](https://quesma.com/blog/baba-is-bench/) |
| NYT Connections (NYT Connections LLM Benchmark (extended version)) | 93.6 | 940 NYT Connections puzzles, each with up to four added trick words | Wordplay, word knowledge and decoys that fit almost-but-not-quite | [NYT Connections LLM Benchmark (Lech Mazur, GitHub)](https://github.com/lechmazur/nyt-connections) |
| StealthMark (StealthMark Intelligence Benchmarks (overall score)) | 50.6 | 8 original tasks: Newton's Mirror, Ground Truth, Latch, Foresight, Mona Lisa, Selfie, Beach and Raccoon | Chained physical reasoning, planning without tools and precise visual construction | [StealthMark: The AI Benchmark Explained (Stealth Models)](https://stealthmodels.com/stealthmark) |
| FrontierPhysics (End-to-End Physics Research Benchmark for AI Agents) | 2.4 | 56 research projects from 42 physicists across nine fields, three trials each (168 rollouts per setup) | Open-ended research projects that took their authors about 270 days on average | [FrontierPhysics: Evaluating Agents for End-to-End Physics Research (arXiv, November 4, 2026)](https://arxiv.org/abs/2611.06910) |
| HoneyBench (HoneyBench v0.2 (Goodhart Labs)) | 78.5 | 20 private honeypot tasks across 7 domains with exploitable evaluators | Agentic alignment and specification gaming | [HoneyBench v0.2](https://goodhartlabs.com/honeybench) |

### Instruction Following

| Benchmark | Value | Tasks | Difficulty | Provenance |
| --- | --- | --- | --- | --- |
| Gray Swan IPI (15 attempts) (Gray Swan Indirect Prompt Injection, attack success at 15 attempts) | 52.7 | Indirect prompt-injection attacks | Adversarial prompt injection | [Gemini 4 Argon evaluation methodology](https://deepmind.google/models/evals-methodology/gemini-4-argon) |

## Lifecycle and limitations log

Lifecycle events the source associates with this model.

No lifecycle event references this model. That is not evidence the model has no lifecycle plan — only that this source published none.

## What this profile does not claim

Values are reproduced exactly as their sources published them, in the units those sources declared; none are converted, interpolated, or averaged across providers. Any field marked unavailable was attempted and not returned. Last attempted fetch for this model's score: 2026-11-21 01:17 UTC.
