K-Exaone vs Qwen3.5-4B

Change models
K-ExaoneLG AI Research
Qwen3.5-4BAlibaba
IndexAll categories, out of 100.
42.3 ±7.9
42.8 ±7.1 best
GapIndex points below the top.
−0.5 · no clear gap
Top
CoverageIndex weight that has results.
75%
80% best
SpeedOutput tokens per second.
—
42/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2025-12
—
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
256K
262K best
AgenticMulti-step tasks with tools.
43.2
64.1 best
CodingCode writing and repair.
37.6
—
ReasoningLogic problems and puzzles.
43.8 best
43.3
MultimodalTasks with images and text.
—
34.6
KnowledgeFacts and expert knowledge.
40.0
40.9 best
MultilingualTasks in many languages.
—
35.7
InstructionTasks with strict prompt rules.
49.0 best
37.6
MathMath problems.
—
40.9

Benchmarks

6 shared
Only K-Exaone: 4Only Qwen3.5-4B: 11
K-Exaone
Qwen3.5-4B
13.9% best124 / 216
9.9%139 / 216
61.3%135 / 211
63.0% best133 / 211
1.1% best113 / 207
0.0%145 / 207
78.3% best99 / 193
77.1%102 / 193
74.3%73 / 143
92.1% best33 / 143
64.7% best53 / 140
52.0%74 / 140