Moonshot AI model product
Kimi K2.6 is Moonshot AI's open-source, native multimodal agentic model focused on state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It scales horizontally to 300 sub-agents executing 4,000 coordinated steps, dynamically decomposing tasks into parallel, domain-specialized subtasks. K2.6 unifies text, image, and video input with thinking and non-thinking modes, supports a 256K context, and powers proactive 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.
Updated Aug 10, 2026. Default version: Kimi K2.6
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Kimi K2.6.
| Claw-Eval | 0.8 | 1 | 13 | 100.0% | C | |
| OJBench | 0.6 | 1 | 9 | 100.0% | C | |
| V* | 1.0 | 1 | 7 | 100.0% | C | |
| WideSearch | 0.8 | 1 | 9 | 100.0% | C | |
| AIME 2026 | 1.0 | 2 | 17 | 93.8% | C | |
| LiveCodeBench v6 | 0.9 | 2 | 53 | 98.1% | C | |
| MathVision | 0.9 | 3 | 32 | 93.5% | C | |
| MCP-Mark | 0.6 | 4 | 8 | 57.1% | C | |
| APEX-Agents | 0.3 | 5 | 7 | 33.3% | C | |
| BabyVision | 0.7 | 5 | 9 | 50.0% | C | |
| HMMT Feb 26 | 0.9 | 5 | 11 | 60.0% | C | |
| SciCode | 0.5 | 5 | 18 | 76.5% | C | |
| DeepSearchQA | 0.8 | 6 | 8 | 28.6% | C | |
| SWE-bench Multilingual | 0.8 | 6 | 34 | 84.8% | C | |
| BrowseComp | 0.9 | 7 | 58 | 89.5% | C | |
| IMO-AnswerBench | 0.9 | 7 | 19 | 66.7% | C | |
| CharXiv-R | 0.9 | 8 | 47 | 84.8% | C | |
| Finance Agent v2 | 0.4 | 11 | 26 | 60.0% | B | |
| SWE-Bench Verified | 0.8 | 12 | 104 | 89.3% | C | |
| FrontierSWE | 0.3 | 13 | 15 | 14.3% | B | |
| MMMU-Pro | 0.8 | 14 | 65 | 79.7% | C | |
| OSWorld-Verified | 0.7 | 14 | 22 | 38.1% | C | |
| Toolathlon | 0.5 | 14 | 31 | 56.7% | C | |
| Terminal-Bench 2.0 | 0.7 | 15 | 49 | 70.8% | C | |
| SWE-Bench Pro | 0.6 | 17 | 44 | 62.8% | C | |
| GPQA | 0.9 | 19 | 233 | 92.2% | C | |
| LiveBench | 0.7 | 25 | 38 | 35.1% | B | |
| Humanity's Last Exam | 0.4 | 39 | 92 | 58.2% | C |
Preference and agent-evaluation signals from published Arena datasets.
| text | chinese | 12 | 1530.8 | 1,873 | N/A | |
| text factuality | japanese | 15 | 1441.6 | 342 | N/A | |
| text style control | chinese | 15 | 1529.2 | 1,873 | N/A | |
| text factuality | chinese | 16 | 1520.2 | 1,601 | N/A | |
| text factuality | math | 16 | 1472.5 | 1,625 | N/A | |
| text factuality | spanish | 16 | 1458.5 | 1,005 | N/A | |
| text factuality | industry mathematical | 17 | 1475.1 | 1,670 | N/A | |
| webdev | image to webdev | 17 | 1520.1 | 3,301 | N/A | |
| text factuality | french | 18 | 1487.9 | 1,196 | N/A | |
| vision style control | chinese | 18 | 1329.8 | 862 | N/A | |
| vision | chinese | 19 | 1349.1 | 862 | N/A | |
| vision | homework | 20 | 1310.0 | 2,149 | N/A | |
| document | overall | 21 | 1450.9 | 11,181 | N/A | |
| text | spanish | 21 | 1463.9 | 1,284 | N/A | |
| vision | humor | 21 | 1280.3 | 539 | N/A | |
| vision style control | homework | 22 | 1302.2 | 2,149 | N/A | |
| vision style control | humor | 22 | 1261.8 | 539 | N/A | |
| text | expert | 23 | 1489.9 | 3,376 | N/A | |
| text | japanese | 23 | 1444.9 | 455 | N/A | |
| text style control | expert | 24 | 1502.6 | 3,376 | N/A | |
| text style control | math | 24 | 1479.4 | 1,921 | N/A | |
| vision | english | 24 | 1285.2 | 6,414 | N/A | |
| document style control | overall | 25 | 1449.5 | 11,181 | N/A | |
| text | korean | 25 | 1426.5 | 712 | N/A | |
| text | math | 25 | 1474.1 | 1,921 | N/A | |
| vision | ocr | 25 | 1290.9 | 11,072 | N/A | |
| vision | overall | 26 | 1281.4 | 15,386 | N/A | |
| vision | diagram | 26 | 1297.5 | 4,040 | N/A | |
| text | industry mathematical | 27 | 1479.0 | 1,998 | N/A | |
| text factuality | expert | 27 | 1497.5 | 2,835 | N/A | |
| webdev | overall | 27 | 1509.7 | 9,261 | N/A | |
| webdev | webdev | 27 | 1509.7 | 9,261 | N/A | |
| webdev | webdev-react | 27 | 1505.1 | 6,764 | N/A | |
| text | industry software and it services | 28 | 1483.9 | 14,915 | N/A | |
| text style control | industry mathematical | 28 | 1485.3 | 1,998 | N/A | |
| vision | creative writing vision | 28 | 1270.1 | 822 | N/A | |
| vision style control | english | 28 | 1265.6 | 6,414 | N/A | |
| text | coding | 29 | 1487.1 | 10,310 | N/A | |
| text style control | japanese | 29 | 1446.0 | 455 | N/A | |
| vision style control | overall | 30 | 1263.4 | 15,386 | N/A | |
| vision style control | diagram | 30 | 1291.7 | 4,040 | N/A | |
| vision style control | ocr | 31 | 1278.5 | 11,072 | N/A | |
| text | overall | 32 | 1454.8 | 37,236 | N/A | |
| text | exclude ties | 32 | 1454.5 | 27,573 | N/A | |
| text | longer query | 32 | 1466.2 | 15,197 | N/A | |
| text | non english | 32 | 1446.8 | 20,191 | N/A | |
| text factuality | industry software and it services | 32 | 1498.5 | 14,668 | N/A | |
| text style control | coding | 32 | 1515.0 | 10,310 | N/A | |
| text | english | 33 | 1458.2 | 17,045 | N/A | |
| text | hard prompts | 33 | 1469.4 | 24,032 | N/A | |
| text style control | industry software and it services | 33 | 1503.8 | 14,915 | N/A | |
| vision style control | creative writing vision | 33 | 1250.6 | 822 | N/A | |
| webdev | webdev-html | 33 | 1495.2 | 1,028 | N/A | |
| text factuality | coding | 34 | 1513.9 | 9,774 | N/A | |
| text | industry business and management and financial operations | 35 | 1445.7 | 7,391 | N/A | |
| text | french | 36 | 1468.6 | 1,491 | N/A | |
| text | hard prompts english | 36 | 1469.1 | 11,617 | N/A | |
| text factuality | longer query | 36 | 1474.1 | 14,772 | N/A | |
| text style control | longer query | 36 | 1475.9 | 15,197 | N/A | |
| text | instruction following | 37 | 1448.8 | 11,994 | N/A | |
| text factuality | industry business and management and financial operations | 37 | 1462.3 | 6,537 | N/A | |
| text factuality | non english | 37 | 1442.8 | 20,163 | N/A | |
| text | german | 38 | 1450.1 | 728 | N/A | |
| text | industry life and physical and social science | 38 | 1467.4 | 6,152 | N/A | |
| text | industry legal and government | 39 | 1458.8 | 2,807 | N/A | |
| text style control | korean | 39 | 1422.3 | 712 | N/A | |
| text style control | spanish | 39 | 1459.8 | 1,284 | N/A | |
| text factuality | overall | 40 | 1455.2 | 37,190 | N/A | |
| text factuality | hard prompts | 40 | 1480.0 | 24,004 | N/A | |
| text factuality | hard prompts english | 40 | 1487.7 | 11,152 | N/A | |
| text factuality | industry entertainment and sports and media | 40 | 1433.0 | 7,014 | N/A | |
| text factuality | industry legal and government | 40 | 1468.0 | 2,287 | N/A | |
| text factuality | instruction following | 40 | 1454.5 | 11,511 | N/A | |
| text | industry entertainment and sports and media | 41 | 1425.2 | 7,541 | N/A | |
| text factuality | industry life and physical and social science | 41 | 1477.3 | 5,337 | N/A | |
| text factuality | industry medicine and healthcare | 41 | 1481.2 | 2,159 | N/A | |
| text factuality | exclude ties | 42 | 1457.2 | 27,537 | N/A | |
| text factuality | russian | 42 | 1459.2 | 3,098 | N/A | |
| text style control | hard prompts | 43 | 1485.0 | 24,032 | N/A | |
| text style control | instruction following | 43 | 1454.5 | 11,994 | N/A | |
| text style control | non english | 43 | 1449.5 | 20,191 | N/A | |
| text | industry writing and literature and language | 44 | 1435.6 | 8,734 | N/A | |
| text factuality | english | 44 | 1463.5 | 17,027 | N/A | |
| text style control | overall | 44 | 1460.7 | 37,236 | N/A | |
| text style control | hard prompts english | 44 | 1486.9 | 11,617 | N/A | |
| text style control | exclude ties | 45 | 1464.3 | 27,573 | N/A | |
| text | creative writing | 46 | 1430.1 | 5,778 | N/A | |
| text factuality | industry writing and literature and language | 46 | 1440.7 | 8,170 | N/A | |
| text style control | german | 46 | 1457.7 | 728 | N/A | |
| text | multi turn | 48 | 1452.6 | 6,490 | N/A | |
| text | russian | 48 | 1444.7 | 3,783 | N/A | |
| text style control | english | 48 | 1467.6 | 17,045 | N/A | |
| text factuality | creative writing | 49 | 1433.7 | 4,964 | N/A | |
| text | polish | 51 | 1440.8 | 737 | N/A | |
| text factuality | multi turn | 52 | 1461.0 | 5,605 | N/A | |
| text style control | french | 53 | 1475.5 | 1,491 | N/A | |
| text style control | industry life and physical and social science | 53 | 1477.6 | 6,152 | N/A | |
| text | industry medicine and healthcare | 54 | 1453.3 | 2,638 | N/A | |
| text style control | industry business and management and financial operations | 54 | 1455.8 | 7,391 | N/A | |
| text style control | creative writing | 56 | 1429.5 | 5,778 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| Privatemode AI | kimi-k2.6 | global | N/A | N/A | 262.1K | |
| Alibaba Token Plan | kimi-k2.6 | global | N/A | N/A | 262.1K | |
| Alibaba Token Plan (China) | kimi-k2.6 | global | N/A | N/A | 262.1K | |
| LLM Gateway | kimi-k2.6 | global | $0.22 | $1.14 | 262.1K | |
| routing.run | kimi-k2.6 | global | $0.275 | $1.1 | 200K | |
| CrofAI | kimi-k2.6 | global | $0.50 | $1.99 | 262.1K | |
| GreenPT | kimi-k2.6 | global | $0.7524 | $4.28 | 262.1K | |
| DigitalOcean | kimi-k2.6 | global | $0.76 | $3.2 | 262.1K | |
| Cortecs | kimi-k2.6 | global | $0.773 | $3.38 | 262.1K | |
| EBCloud | Kimi-K2.6 | global | $0.9286 | $3.86 | 262.1K | |
| Alibaba (China) | kimi-k2.6 | global | $0.929 | $3.86 | 262.1K | |
| OpenCode Go | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| OpenCode Zen | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Azure Cognitive Services | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Moonshot AI (China) | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Auriko | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Azure | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| AIHubMix | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Charm Hyper | kimi-k2.6 | global | $0.95 | $4 | 262K | |
| Moonshot AI | kimi-k2.6 | global | $0.95 | $4 | 262.1K | |
| Wafer | Kimi-K2.6 | global | $1.14 | $4.8 | 262.1K | |
| UnoRouter | kimi-k2.6 | global | $1.27 | $5.34 | 262.1K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Kimi K2.6 | 82.8 | 1T | 262.1K | 262.1K | Yes | Modified MIT License |
A concise description based on the published model registry.
Kimi K2.6 is Moonshot AI's open-source, native multimodal agentic model focused on state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It scales horizontally to 300 sub-agents executing 4,000 coordinated steps, dynamically decomposing tasks into parallel, domain-specialized subtasks. K2.6 unifies text, image, and video input with thinking and non-thinking modes, supports a 256K context, and powers proactive 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Kimi K2.6.
Kimi K2.6's default version was released on Apr 20, 2026.
Kimi K2.6's official API price is $0.95 per million input tokens and $4 per million output tokens via Moonshot AI. The lowest tracked third-party offer starts at $0.22 input and $1.14 output via LLM Gateway.
Kimi K2.6 is published under Moonshot AI in the model registry.
The default version has a 262.1K token context window.
Yes. The default version is marked as open weight under Modified MIT License.
23 published provider offerings are linked to the default version.
Nearby ranked alternatives include GPT-5.6-Terra, Gemini 3.1 Pro, Claude Opus 4.8.