HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Kimi-K2-Thinking

unsloth/Kimi-K2-Thinking

kind
original
quant
bf16 (16-bit)
family
kimi
modality
text-generation
root model
moonshotai/Kimi-K2-Thinking
provider
unsloth
license
other
language
—
context
256K
library
transformers
vRAM (GB)
—
size on disk
—
downloads
71
likes
4
created
—
last modified
2026-01-27T06:54:03+00:00 (2026)
huggingface
https://huggingface.co/unsloth/Kimi-K2-Thinking
tags
transformerssafetensorsdeepseek_v3text-generationunslothconversationalcustom_codebase_model:moonshotai/Kimi-K2-Thinkingbase_model:quantized:moonshotai/Kimi-K2-Thinkinglicense:othertext-generation-inferenceendpoints_compatiblecompressed-tensorsregion:us

Benchmark scores (11)

benchmarkcategoryscorecontextconfevidence
AIME 2025math94.5**AIME25**medium| **AIME25** | no tools | 94.5 | 94.6 | 87.0 | 51.0 | 89.3 | 91.7 |
MMLUknowledge94.4**MMLU-Redux**medium| **MMLU-Redux** | no tools | 94.4 | 95.3 | 95.6 | 92.7 | 93.7 |
MMLU-Proknowledge84.6**MMLU-Pro**medium| **MMLU-Pro** | no tools | 84.6 | 87.1 | 87.5 | 81.9 | 85.0 |
GPQAreasoning84.5**GPQA**medium| **GPQA** | no tools | 84.5 | 85.7 | 83.4 | 74.2 | 79.9 | 87.5 |
LiveCodeBenchcoding83.1**LiveCodeBenchV6**medium| **LiveCodeBenchV6** | no tools | 83.1 | 87.0* | 64.0* | 56.1* | 74.1 |
SWE-bench Verifiedswe71.3**SWE-bench Verified**medium| **SWE-bench Verified** | w/ tools | 71.3 | 74.9 | 77.2 | 69.2 | 67.8 |
BrowseCompagentic62.3**BrowseComp-ZH**medium| **BrowseComp-ZH** | w/ tools | 62.3 | 63.0* | 42.4* | 22.2 | 47.9 |
BrowseCompagentic60.2**BrowseComp**medium| **BrowseComp** | w/ tools | 60.2 | 54.9 | 24.1 | 7.4 | 40.1 |
Terminal-Benchagentic-coding47.1**Terminal-Bench**medium| **Terminal-Bench** | w/ simulated tools (JSON) | 47.1 | 43.8 | 51.0 | 44.5 | 37.7 |
SciCodecoding44.8**SciCode**medium| **SciCode** | no tools | 44.8 | 42.9 | 44.7 | 30.7 | 37.7 |
Humanity's Last Examknowledge23.9**HLE (Text-only)**medium| **HLE (Text-only)** | no tools | 23.9 | 26.3 | 19.8* | 7.9 | 19.8 | 25.4 |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.