[whatmodel.sh](../../README.md) / A.X K2 (688B-A33B) # A.X K2 (688B-A33B) SK Telecom · released 2026-07 · 691.69B params (33B active) · apache-2.0 **score 68.1** ±1.3 · #98 of 732 · from 7 benchmark results across 1 source · † no independently run results verified Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks A.X K2 (688B-A33B) among every model on that task or test, worst to best.}

task    vs the others         rank  counts  tests
coding  ▁▁▁▃▅▇█▇▅●▃▁   #78 of 469*     17%      2
math    ▁▁▃▅█▆▄█●▂▁▁   #49 of 431*    7.5%      1

coding · #78 of 469 · 17% of its score

Beats [GPT-5.2](gpt-5-2.md), [GPT-5](gpt-5.md), [Grok 4.20 (0309 reasoning)](grok-4-20.md) here, though they rank higher overall. Trails [Llama 3.1 Nemotron Ultra 253B v1](llama-3-1-nemotron-ultra-253b.md), [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty        counts  why
SciCode (level unspecified) (make…   41.0%  ▄·····▄●▄▄·▄ beats 2/6    hard · avg 16%      8.9%  at its level
LiveCodeBench v6 (Feb–May 2025) (…   84.0%  ▄··█····▄●·█ beats 4/6    medium · avg 47%    8.4%  typical

math · #49 of 431 · 7.5% of its score

Beats [Grok 4.7](grok-4-7.md), [GLM-5.3-Flash](glm-5-3-flash.md), [LongCat 2.0 (1.6T-A48B)](longcat-2-0.md) here, though they rank higher overall.

test               result  vs the others in its run  difficulty        counts  why
AIME 2026 (maker)   97.1%  ▄·▄·▄▄··█··● beats 6/6    medium · avg 57%    7.5%  sharp · maxed out · noisy

other tests (4)

test                                result  vs the others in its run  difficulty            counts  why
GPQA Diamond (maker)                 85.6%  ▄·····●█▄▄·▄ beats 1/6    hard · avg 44%           35%  sharp
Global-MMLU-Lite (maker)             87.6%  ▃······●··█▃ beats 2/6    medium · avg 61%         16%  typical
Humanity's Last Exam (maker)         27.8%  ▂········●·▂ beats 3/6    very hard · avg 5.4%     15%  at its level
IFBench (aggregation unspecified)…   75.9%  ▃······●▃▃·▃ beats 2/6    medium · avg 45%        9.2%  typical

## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its *counts* column above; here it is added up by who ran the tests. [Method](../method.md)

source                                    results  share of the score
A.X K2 Technical Report (submitted / ma…        7   99% ▇▇▇▇▇▇▇▇▇▇▇▇
prior: pull toward the middle                        1% ▇
weighted average                                   = 68.1

_{Licences: A.X K2 Technical Report ([CC BY 4.0](http://creativecommons.org/licenses/by/4.0/))} ## How to run it ### Other ways to run it 1. Run it as published (BF16, needs ~1383.4 GB, NVIDIA only, from [huggingface.co/skt](https://huggingface.co/skt/A.X-K2/tree/2287ca456927eed33b899c404c0b2ffaf17aa09f)) ```sh vllm serve skt/A.X-K2 --tensor-parallel-size 8 ``` Model repository and documentation: [huggingface.co/skt](https://huggingface.co/skt/A.X-K2) [interactive version](https://whatmodel.sh/model/ax-k2) · [JSON](https://whatmodel.sh/api?format=json&all=1)