← Back to leaderboard
Head-to-head
Phi 3.5 MoE 42B vs openai/gpt-oss-20b
openai/gpt-oss-20b wins+23.0 pts (+33%)on the balanced composite.
Category-by-category
Higher bar wins each row.
Phi 3.5 MoE 42B
Category
openai/gpt-oss-20b
-28.571.5
Code
100.0+28.5
-4.170.9
Reason
75.0+4.1
-16.271.3
Tool Use
87.5+16.2
-26.773.3
RAG
100.0+26.7
-36.163.2
Speed
99.3+36.1
Phi 3.5 MoE 42B
- Provider
- microsoft
- Released
- 2024-08-20
- Context
- 131.072K
- Lab
- Community
openai/gpt-oss-20b
- Provider
- local
- Released
- Context
- 0K
- Lab
- Community