DeepSeek-R1 vs Gemma 4 12B

Change models
DeepSeek-R1DeepSeek
IndexAll categories, out of 100.
38.9 ±5.0
39.0 ±4.0 best
GapIndex points below the top.
−0.1 · no clear gap
Top
CoverageIndex weight that has results.
85%
100% best
SpeedOutput tokens per second.
18/s
—
EloLMArena rating and rank.
1373
—
ReleasedMonth of release.
2025-01
2026-06
Input / 1MUS dollars per 1M input tokens.
$0.70
—
Output / 1MUS dollars per 1M output tokens.
$2.50
—
1K chats2K tokens in, 500 out, each.
$2.65
—
1K documents20K tokens in, 1K out, each.
$16.50
—
1K agent tasks200K tokens in, 20K out, each.
$190.00
—
ContextMaximum tokens in one request.
64K
256K best
AgenticMulti-step tasks with tools.
27.6
30.5 best
CodingCode writing and repair.
42.6
43.5 best
ReasoningLogic problems and puzzles.
39.8
41.5 best
MultimodalTasks with images and text.
—
36.2
KnowledgeFacts and expert knowledge.
43.5 best
41.5
MultilingualTasks in many languages.
—
29.7
InstructionTasks with strict prompt rules.
28.5
57.7 best
MathMath problems.
43.8 best
41.1

Benchmarks

8 shared
Only DeepSeek-R1: 26Only Gemma 4 12B: 20
DeepSeek-R1
Gemma 4 12B
15.8% best123 / 222
15.7%126 / 222
55.7%146 / 217
63.7% best137 / 217
1.4% best113 / 214
0.0%150 / 214
81.3% best94 / 199
75.3%115 / 199
36.5% best107 / 149
36.3%108 / 149
39.6%105 / 146
73.5% best28 / 146
71.5%87 / 141
78.8% best71 / 141
84.0% best27 / 75
77.2%45 / 75