PipelineScore
← Back to leaderboard
Head-to-head

Dolphin 2.9 Llama 3 8B vs DeepSeek V3 671B-A37B

DeepSeek V3 671B-A37B wins+17.7 pts (+28%)on the balanced composite.
cognitivecomputations
Dolphin 2.9 Llama 3 8B
64.3FEEDER
deepseek
Winner
DeepSeek V3 671B-A37B
82.0MAINLINE

Category-by-category

Higher bar wins each row.

Dolphin 2.9 Llama 3 8B
Category
DeepSeek V3 671B-A37B
-18.462.9
Code
81.3+18.4
-15.866.4
Reason
82.3+15.8
-22.561.7
Tool Use
84.3+22.5
-18.366.8
RAG
85.1+18.3
-13.066.8
Speed
79.7+13.0
Dolphin 2.9 Llama 3 8B
Provider
cognitivecomputations
Released
2024-05-20
Context
8.192K
Lab
Community
DeepSeek V3 671B-A37B
Provider
deepseek
Released
2024-12-26
Context
128K
Lab
Community
Dolphin 2.9 Llama 3 8B vs DeepSeek V3 671B-A37B · PipelineScore