Skip to content
Local AI Wizard
Your machine
Theme

Kimi K2.6 · thinking

Moonshot AI · moonshotai/Kimi-K2.6

Q4 file 583.7 GB (UD-Q4_K_XL) · Weights license: Publisher terms: Modified MIT

LiveBench · overall70.5 / 100#17 of 22 open models · 2026-06-25
Memory needed~750 GBQ4 weights + working room, before the machine's reserve
Parameters1,000 Bmanufacturer specification · 32B active
Context262,144 tokensPublished for this model

Quality

Quality · Overall

LiveBench 2026-06-25 · 0–100 · higher is better · this model outlined, #17 of 22

  • Color = publisher
  • Memory under review

Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)

No result in LMArena 2026-10-22 Not in LMArena 2026-10-22

LiveBench categories

This model's mean per category · 0–100

79.4ReasoningReasoning78.6CodingCoding46.9Agentic CodingAgentic Coding84.3MathematicsMathematics65.1Data AnalysisData Analysis75.1LanguageLanguage64.4IFIF

Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)

Reported by the publisher

33 results · Moonshot AI's model card · read 2026-10-25
HLE-Full (w/ tools)
54.0
BrowseComp
83.2
BrowseComp (Agent Swarm)
86.3
DeepSearchQA (f1-score)
92.5
DeepSearchQA (accuracy)
83.0
WideSearch (item-f1)
80.8
Toolathlon
50.0
MCPMark
55.9
Every reported result33
Claw Eval (pass^3)
62.3
Claw Eval (pass@3)
80.9
APEX-Agents
27.9
OSWorld-Verified
73.1
Terminal-Bench 2.0 (Terminus-2)
66.7
SWE-Bench Pro
58.6
SWE-Bench Multilingual
76.7
SWE-Bench Verified
80.2
SciCode
52.2
OJBench (python)
60.6
LiveCodeBench (v6)
89.6
HLE-Full
34.7
AIME 2026
96.4
HMMT 2026 (Feb)
92.7
IMO-AnswerBench
86.0
GPQA-Diamond
90.5
MMMU-Pro
79.4
MMMU-Pro (w/ python)
80.1
CharXiv (RQ)
80.4
CharXiv (RQ; w/ python)
86.7
MathVision
87.4
MathVision (w/ python)
93.2
BabyVision
39.8
BabyVision (w/ python)
68.5
V (w/ python)
96.9

Values as printed on the card, with the publisher's own settings (card section: 3. Evaluation Results); column “Kimi K2.6”. Not comparable across publishers and not used in any ranking. Model card

Run it

Run it

File UD-Q4_K_XL · 583.7 GB · 14 parts · unsloth/Kimi-K2.6-GGUF · What is Q4?

llama-server -hf unsloth/Kimi-K2.6-GGUF:UD-Q4_K_XL
  1. Install llama.cpp
  2. Run the command above
  3. Ask at localhost:8080: Explain what a context window is, in two sentences.

Quality × memory

Quality × memory

LiveBench 2026-06-25 overall vs. Q4 planning memory (log scale) · 18 open models

  • This model
  • Color = publisher
60708081632641282565121,0242,0484,096Memory needed · Q4 · model only (GB, log)LiveBenchSmaug AgenticQwen3.8 MaxDeepSeek V4 Pro · 0813Smaug MiniKimi K2.6 · thinking

Memory = the model's own figure: parameters × 0.625 bytes + 20% working room, before any machine's reserve; with a chosen machine, the total there with its reserve (25% of memory, 2 to 8 GB; 2 GB per graphics card). An estimate, not measured VRAM. LiveBench 2026-06-25

Runs on

More memory than any profile in the catalog

Even the largest profile (8 × 80 GB, aggregated) is below this planning estimate. Multi-node setups are outside this catalog.

Kimi K2.6 · thinking is a text model from Moonshot AI with 1,000B parameters, 32B of them active per token (mixture of experts). It reads up to 262,144 tokens of context. Its Q4 weights plus working room come to about 750 GB, more than the fast memory of every reference machine. We collected 4 downloadable GGUF files from one publisher; the most downloaded GGUF source is unsloth's. It scores 70.5 overall in LiveBench 2026-06-25.

On the reference machines

Kimi K2.6 · thinking on the reference machines
Where Q4 sits
Mac mini M2 · 8 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac mini M4 · 16 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~754 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac mini M4 · 24 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~756 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac mini M4 · 32 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~758 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac mini M5 Pro · 64 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~758 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac Studio M5 Max · 128 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~758 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
Mac Studio · M5 Ultra · 36-core CPU / 80-core GPU · 512 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~758 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
RTX 4060 · 8 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
RTX 3060 · 12 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
RTX 5080 · 16 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
RTX 4090 · 24 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
RTX 5090 · 32 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~752 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column
DGX H100 · 8 × 80 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column~766 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column

The file column uses the site's quant ladder (Q8_0 to Q3_K_M, MLX 8 to 4-bit on a Mac): the largest collected file whose size × 1.2 plus the context cache and the machine's reserve fits the fast memory. Planning estimates, not measured runs.

Plan another quantization or context

Consider it for

Inspect this thinking variant when comparing larger reasoning checkpoints using the same benchmark snapshot.

Check first

The model card lists a vision encoder separately; experimental video support is API-only, not a claimed local feature.

Versions to download

Files you can download for this checkpoint. The section opens on the most downloaded collected GGUF source, unsloth, at UD-Q4_K_XL. Sizes are the Hub's published bytes. What is Q4?

GGUF · the only format collected

unsloth

unsloth/Kimi-K2.6-GGUF · UD-Q4_K_XL/Kimi-K2.6-UD-Q4_K_XL-00001-of-00014.gguf · 583.7 GB

Run this file

llama-server -hf unsloth/Kimi-K2.6-GGUF:UD-Q4_K_XL
lms get https://huggingface.co/unsloth/Kimi-K2.6-GGUF
  • Ollama: Ollama's documented hf.co form covers single-file GGUFs; this file has 14 parts
  • mlx-lm: mlx-lm runs MLX repositories; pick the MLX version on a Mac

Quality scores on this page belong to the original checkpoint; these files are not scored or ranked. File sizes read 2026-10-23 · download counts read 2026-10-17 · commands: Hugging Face Hub docs: Use Ollama with any GGUF model · llama.cpp server README (-hf <user>/<model>[:quant]) · LM Studio docs: lms get with a Hugging Face URL · mlx-lm README (mlx_lm.generate --model)

Fine-tunes, adapters and other versions19 fine-tunes
Quantized versions
45
Fine-tunes
19
Adapters · merges
9 · 2

Browse them on the Hugging Face model tree · read 2026-10-23

Uncensored and abliterated versions5

Some community uploads remove the model's safety refusals. Quality may drop, the original publisher does not support them, claims about them are the uploader's, and the original license still applies. This site never recommends or ranks them.

5 derivative repositories have “uncensored” or “abliterated” in their name (read 2026-10-23).

Measured speed

No published measurement for this checkpoint

Tokens per second are shown only when a source publishes them with hardware, build and setup. They are never estimated.

How to run

~750 GB of Q4 weights plus working room, before the machine's reserve. Formula: parameters × 0.625 bytes, plus 20% working allowance; each machine then adds its system reserve: 25% of unified or system memory, 2 to 8 GB, or 2 GB per GPU on dedicated cards. This is neither measured VRAM nor checkpoint download size.

Q4 planning estimate only. Check checkpoint format, runtime support, KV cache, context, batch size and offload before choosing hardware.

Evidence level: Generic estimate only; no execution reproduced by this site.

Publisher sampling defaults
— generation_config.json states no sampling values for this checkpoint
Chat template markers
tool calls · thinkingmarkers found in the template, not a test of the feature · read 2026-10-23source
Before downloading
  1. Check the exact checkpoint and license linked here.
  2. Confirm a compatible runtime and precision or quantization for your platform.
  3. Allow for context, batch size and any extra media components beyond the weights.

Technical details

What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.

Checkpoint / version
moonshotai/Kimi-K2.6
Task and modality
language and reasoning · text + image
Count and basis
1,000B · manufacturer specification · 32B active
Context
262,144 tokens published for this model
Architecture
MoE
Runtime and precision
No local execution verified.
Commercial use
"Modified MIT" names different modified-MIT texts (Kimi's and MiniMax's, with different conditions), so the name alone does not settle commercial use
Every stored field10 fields
Revision checked
7eb5002f6aadc958aed6a9177b7ed26bb94011bb
Architecture
KimiK25ForConditionalGeneration
Library
transformers
Hub task tag
image-text-to-text
Formats
safetensors
Repository created
2026-04-14creation is not the model's release
Repository updated
2026-05-19a commit is not a release date
Hub downloads (30 days) · likes
353,428 · 1,636this repository only · read 2026-10-25
Read from the Hub page
task, transformers class, task tags, license name, file formatsHub overview and file listing · read 2026-10-18
Catalog state
Hub config or weights found; no inventory benchmark link

The model card lists a 400M vision encoder separately. Experimental video support is API-only, so it is not claimed for local inference.

Sources

Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.

Catalog reviewed: 2026-09-30. Benchmark release: 2026-06-25; snapshot collected: 2026-09-26.