Kimi-VL-A3B-Thinking-2506

Home/Models/moonshotai/Kimi-VL-A3B-Thinking-2506

moonshotai/Kimi-VL-A3B-Thinking-2506

kind: original
quant: bf16 (16-bit)
family: kimi
modality: image-text-to-text
root model: moonshotai/Kimi-VL-A3B-Instruct
provider: moonshotai
license: mit
language: —
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 7,466
likes: 371
created: —
last modified: 2026-01-30T04:53:11+00:00 (2026)
huggingface: https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking-2506
tags: transformerssafetensorskimi_vlfeature-extractionimage-text-to-textconversationalcustom_codearxiv:2504.07491base_model:moonshotai/Kimi-VL-A3B-Instructbase_model:finetune:moonshotai/Kimi-VL-A3B-Instructlicense:miteval-resultsregion:us

benchmark	category	score	context	conf	evidence
MATH	math	91.8	MATH	medium	`\| MATH \| 91.8 \| 82.2 \| 89.0 \|`
MMBench	multimodal	83.1	MMBench-EN-v1.1 (Acc)	medium	`\| MMBench-EN-v1.1 (Acc) \| 83.1 \| 83.2 \| 74.6 \| 82.9 \| 76.0 \| 84.4 \|`
MMLU	knowledge	82.0	MMLU	medium	`\| MMLU \| 82.0 \| 78.4 \| 76.9 \|`
MathVista	multimodal	80.1	MathVista\_MINI (Pass@1)	medium	`\| MathVista\_MINI (Pass@1) \| 80.1 \| 74.7 \| 74.8 \| 59.8 \|`
MMMU	multimodal	69.1	MMMU (val, Pass@1)	medium	`\| MMMU (val, Pass@1) \| 69.1 \| 58.6 \| 59.6 \| 57.0 \| 61.7 \| 64.0 \|`
MMLU-Pro	knowledge	68.5	MMLU-Pro	medium	`\| MMLU-Pro \| 68.5 \| 68.8 \| 67.5 \|`
MathVista	multimodal	63.8	MathVista_MINI (Pass@1)	medium	`\| MathVista_MINI (Pass@1) \| 63.8 \| 68.0 \| 56.1 \| 68.6 \| 71.7 \| 80.1 \|`
MMMU	multimodal	51.7	MMMU-Pro (Pass@1)	medium	`\| MMMU-Pro (Pass@1) \| 51.7 \| 38.1 \| 32.1 \| 36.0 \| 43.2 \| 46.3 \|`
LongBench	long-context	42.8	MMLongBench-DOC (Acc)	medium	`\| MMLongBench-DOC (Acc) \| 42.8 \| 29.6 \| 21.3 \| 35.1 \| 32.5 \| 42.1 \|`
GPQA Diamond	reasoning	42.3	GPQA-Diamond	medium	`\| GPQA-Diamond \| 42.3 \| 46.0 \| 46.0 \|`
OSWorld	agentic	31.5	OSWorld-G (Acc)	medium	`\| OSWorld-G (Acc) \| - \| 31.5 \| — \| 41.6 \| — \| 52.5 \|`
MATH	math	30.4	MATH-Vision (Pass@1)	medium	`\| MATH-Vision (Pass@1) \| 30.4 \| 25.0 \| 32.1 \| 21.7 \| 36.8 \| 56.9 \|`

none scraped.