Inference
EN
SK Telecom

A.X K2

A.X K2 is a reasoning model from SK Telecom in the A.X family. 10 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-07
RankedEnough results for an index.
IndexAll categories, out of 100.51.9±6.6Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

691.7B parameters
16-bit
1453.6 GB
8-bit
772.7 GB
4-bit
441.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted46.3159 / 218
CodingCode writing and repair.2 counted51.3130 / 209
ReasoningLogic problems and puzzles.2 counted53.395 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.2 counted54.599 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted60.0—
MathMath problems.1 counted55.2—

Results

10 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
AIME 2026Math97.1%5 / 34
Critical Physics TasksReasoning8.9%72 / 214
Artificial Analysis GPQA DiamondKnowledge85.7%72 / 199
Artificial Analysis Humanity's Last ExamKnowledge29.6%82 / 222
LiveCodeBench v6Coding84.0%23 / 59
Instruction Following BenchmarkInstruction75.9%27 / 57
Scientific Code BenchmarkCoding41.0%20 / 35
GDPval-AA normalizedAgentic23.0%78 / 133
Artificial Analysis Long Context ReasoningReasoning66.0%133 / 217
GPQA Diamond · BenchLMKnowledge85.6%57 / 84
Humanity's Last ExamKnowledge27.8%54 / 77
Artificial Analysis SciCodeCoding41.0%83 / 114
Terminal-Bench 2.1 · BenchLMAgentic36.0%80 / 86
Apex ShortlistMath88.6%—
ApexMath45.8%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index51.3 → 51.9
23 Oct 202612 Nov 2026

Questions

How good is A.X K2?
Its index is 51.9 out of 100. That is place 133 of 283 ranked models.
What is A.X K2 best at?
Knowledge: 54.5 out of 100, place 99 of 268.
What is the context window of A.X K2?
262K tokens.
Is A.X K2 open source?
Yes. Its weights are published.
When was A.X K2 released?
July 2026.

Badge

A.X K2 on Inference360: place 133 of 283, index 51.9
HTML
<a href="https://inference360.ai/models/a-x-k2"><img src="https://inference360.ai/badges/a-x-k2.svg" alt="A.X K2 on Inference360: place 133 of 283, index 51.9" height="20"></a>
Markdown
[![A.X K2 on Inference360: place 133 of 283, index 51.9](https://inference360.ai/badges/a-x-k2.svg)](https://inference360.ai/models/a-x-k2)

Compare with

GLM-5-Turbo51.9Claude Opus 4.552.0Gemini 3.5 Flash-Lite52.0

More from SK Telecom

All models
A.X seriesUnranked