DeepSeek V3 0324 vs Nemotron 3.5 Lightning 30B A3B NVFP4

Change models
IndexAll categories, out of 100.
37.1 ±5.0 best
37.0 ±6.8
GapIndex points below the top.
Top
−0.1 · no clear gap
CoverageIndex weight that has results.
90% best
75%
SpeedOutput tokens per second.
—
—
EloLMArena rating and rank.
1375
—
ReleasedMonth of release.
2025-03
2026-08
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.
128K
1M best
AgenticMulti-step tasks with tools.
34.0 best
30.2
CodingCode writing and repair.
40.9 best
38.7
ReasoningLogic problems and puzzles.
36.0
38.9 best
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
37.6
39.5 best
MultilingualTasks in many languages.
40.1
—
InstructionTasks with strict prompt rules.
27.3
57.1 best
MathMath problems.
38.0
—

Benchmarks

7 shared
Only DeepSeek V3 0324: 15Only Nemotron 3.5 Lightning 30B A3B NVFP4: 12
DeepSeek V3 0324
Nemotron 3.5 Lightning 30B A3B NVFP4
4.7%183 / 222
10.6% best143 / 222
40.7%167 / 217
49.2% best159 / 217
0.0%150 / 214
0.0%150 / 214
65.5%144 / 199
74.3% best122 / 199
24.3% best108 / 164
14.7%153 / 164
0.0%101 / 133
6.9% best93 / 133
39.0% best91 / 114
32.1%102 / 114