DeepSeek-R1 vs Grok 3 [Beta]

Change models
DeepSeek-R1DeepSeek
IndexAll categories, out of 100.
38.8 ±5.0
39.1 ±8.5 best
GapIndex points below the top.
−0.3 · no clear gap
Top
CoverageIndex weight that has results.
85% best
60%
SpeedOutput tokens per second.
18/s
—
EloLMArena rating and rank.
1373
—
ReleasedMonth of release.
2025-01
2025-02
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
128K best
AgenticMulti-step tasks with tools.
28.2
—
CodingCode writing and repair.
42.3 best
34.7
ReasoningLogic problems and puzzles.
39.7 best
34.6
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
43.4
46.8 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
28.8
—
MathMath problems.
43.5 best
37.9

Benchmarks

19 shared
Only DeepSeek-R1: 15Only Grok 3 [Beta]: 7
DeepSeek-R1
Grok 3 [Beta]
71.7%114 / 220
75.8% best106 / 220
53.3%121 / 203
55.6% best119 / 203
67.3%133 / 149
82.6% best67 / 149
71.5%87 / 144
75.4% best79 / 144
TaxEval v2Finance
72.3%66 / 143
75.9% best10 / 143
70.2% best88 / 142
52.9%114 / 142
1.3% best108 / 137
0.0%126 / 137
MMLU ProKnowledge
83.2% best73 / 135
79.9%91 / 135
15.8% best116 / 132
5.5%125 / 132
CorpFin v2Finance
54.1%97 / 131
59.7% best74 / 131
0.3% best104 / 127
0.0%107 / 127
AIMEMath
74.0% best52 / 95
58.8%58 / 95
MedQAHealthcare
90.8% best44 / 95
83.8%63 / 95
93.1% best13 / 88
88.7%17 / 88
MGSMMultilingual
92.3% best20 / 74
91.3%28 / 74
84.0% best24 / 73
79.9%36 / 73
92.2% best18 / 59
89.8%27 / 59
79.8% best13 / 28
52.2%23 / 28
30.1%5 / 15
43.6% best3 / 15