Back to Benchmarks

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) vs DeepSeek R1 Distill Llama 70B

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Anthropic

vs

DeepSeek R1 Distill Llama 70B

DeepSeekvia DeepInfra

Tied across all benchmarks

Index Scores

51.7
+35.7
Intelligence
16.0
50.9
+39.5
Coding
11.4
0.0
Math
+53.7
53.7

Benchmarks

0.0
MMLU Pro
+79.5
79.5
87.5
+47.3
GPQA
40.2
30.0
+23.9
HLE
6.1
0.0
LiveCode
+26.6
26.6
46.8
+15.6
SciCode
31.2
0.0
MATH-500
+93.5
93.5
0.0
AIME
+67.0
67.0

Speed

42
TPS
42
Shared
Winning delta

crafted by bart stefanski