BlogLLM Benchmark Scores as Published, Side by Side
Kimi K2.6: Benchmark Scores as Published
Every result the catalog holds for Kimi K2.6: 78 tests in 14 areas, from 10 pages. 38 printed by Moonshot AI itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer Moonshot AIReleased 2026-04-20Context window 256KInput Image, videoOutput TextWeights Open (Modified MIT License)Data version 2026103001, as of 2026-10-30
Knowledge Work
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 80.2% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 5 more pages | Harness: in-house SWE-agent variant |
| SWE-Bench Pro | 58.6% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 4 more pages | Harness: in-house SWE-agent variant |
| Terminal-Bench 2.1 | 67.2% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15; printed on 1 more page | - |
| SWE-bench Multilingual | 76.7% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | Harness: in-house SWE-agent variant |
| Terminal-Bench 2.0 | 66.7% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | Harness: Terminus-2 |
| NL2Repo-Bench | 42.8 | MiniMax, MiniMax M3checked 2026-10-15 | Effort: thinking |
| ProgramBench | 48.3% | Moonshot AI, Kimi K2.7 Code model cardchecked 2026-10-15 | Context: 262144 |
| Kimi Code Bench 2.0 | 50.9% | Moonshot AI, Kimi K2.7 Code model cardchecked 2026-10-15 | Context: 262144 |
ML Engineering
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MLS-Bench-Lite | 26.7 | Moonshot AI, Kimi K2.7 Code model cardchecked 2026-10-15 | Context: 262144 |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 90.5% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 3 more pages | - |
| SciCode | 52.2% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | - |
| IMO-AnswerBench | 86% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | - |
| AIME 2026 | 96.4% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | - |
| HMMT February 2026 | 92.7% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | - |
| CritPt | 9.1% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | Tools: none |
| MathArena Apex Shortlist | 75.5% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15; printed on 1 more page | Effort: thinking |
| MathArena Apex | 24% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15 | Effort: thinking |
| GPQA | 91% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | Tools: none |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AA-LCR | 70.2% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OSWorld-Verified | 73.1% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | - |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CharXiv Reasoning (no tools) | 80.4% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | Tools: none |
| MMMU-Pro | 79.4% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | Tools: none |
| MathVision | 87.4% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: none |
| CharXiv Reasoning (with tools) | 86.7% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | Tools: python |
| BabyVision (with tools) | 68.5% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: python |
| MMMU-Pro (with tools) | 80.1% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: python |
| BabyVision (no tools) | 39.8% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: none |
| MathVision (with tools) | 93.2% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: python |
| V* Bench (with tools) | 96.9% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: python |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 34.7% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 3 more pages | Tools: none |
| MMLU-Pro | 87.1% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15; printed on 1 more page | Effort: thinking |
| Humanity's Last Exam (with tools) | 54% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 3 more pages | Tools: search, code interpreter, browsing; Context: 262144 |
| AA-Omniscience Accuracy | 35.5% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
| SimpleQA Verified | 36.9% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15; printed on 1 more page | Effort: thinking |
| AA-Omniscience Index | 6 | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | - |
| AA-Omniscience Non-Hallucination | 67.1% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
| Chinese SimpleQA | 75.9% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15 | Effort: thinking |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LiveCodeBench v6 | 89.6% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 2 more pages | - |
| LiveCodeBench (version not stated) | 89.6% | DeepSeek, DeepSeek-V4-Pro (preview) model cardchecked 2026-10-15 | Effort: thinking |
| IOI 2025 (total points) | 585 | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
| OJBench (Python) | 60.6% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | - |
Tool Use and Workflows
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 83.2% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 4 more pages | Tools: browsing with context management |
| 86.3% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: browsing, agent swarm | |
| DeepSearchQA (accuracy) | 83% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: search, code interpreter, browsing |
| DeepSearchQA (F1) | 92.5% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: search, code interpreter, browsing |
| WideSearch (item-F1) | 80.8% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15 | Tools: search, code interpreter, browsing |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| IFBench | 73.7 | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15; printed on 1 more page | - |
| MultiChallenge | 63.1% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MMLU-ProX | 85% | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
| WMT24++ (XCOMET-XXL) | 84.5 | NVIDIA, NVIDIA Nemotron 3 Ultra 550B-A55B model cardchecked 2026-10-15 | - |
| Global-MMLU-Lite | 88.4% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | - |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| FORTRESS (Adversarial) | 65.6% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | - |
| FORTRESS (Benign) | 97.2% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | - |
| StrongREJECT | 99.8% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | - |
| B3 AI Security Benchmark (attack resistance) | 83.8% | Mistral AI, Mistral Large 4 (preview)checked 2026-10-15 | - |
| Cyber refusal rate (average of JailbreakBench, StrongREJECT, AgentHarm) | 90.7% | Mistral AI, Mistral Large 4 (preview)checked 2026-10-15 | - |
| KORA Benchmark (0 to 2) | 1.6 | Mistral AI, Mistral Large 4 (preview)checked 2026-10-15 | - |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked. Cost per task is measured by Vals AI.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Input | $0.95 | Moonshot AI, Kimi API pricing (international platform)checked 2026-10-19 | - |
| Cached input | $0.16 | Moonshot AI, Kimi API pricing (international platform)checked 2026-10-19 | - |
| Output | $4 | Moonshot AI, Kimi API pricing (international platform)checked 2026-10-19 | - |
| Cost per taskVals AI, USD, lower is better | $1.93 | Vals AI, Vibe Code Bench v1.1 leaderboardchecked 2026-10-27 | Harness: OpenHands |
More from Moonshot AI
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.