DeepSeek Math 7B Instruct
Open Source
Capabilities
VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode Execution
Benchmark Scores(4)
| Benchmark | Score | Version | Source |
|---|---|---|---|
| Google-Proof Q&A | 49.2 | diamond | Open LLM Leaderboard |
| HellaSwag | 90.1 | 10-shot | Open LLM Leaderboard |
| HumanEval | 78.9 | pass@1 | Open LLM Leaderboard |
| Massive Multitask Language Understanding | 75.9 | 5-shot | Open LLM Leaderboard |
Specifications
FamilyDeepSeek Math
Released2024-02-05
Parameters7B
ArchitectureDecoder Only
Specializationmath
Trainingfinetuning
Fine-tuninginstruct