DeepSeek R1 Distill Llama 8B

DeepSeek · deepseek/deepseek-r1-distill-llama-8b

Provider	DeepSeek
Family	llama
Type	llm-reasoning
Status	active
Released	2025-01-20

Reported benchmark scores

These scores carry one collection date for the whole card (2026-04) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.

Benchmark	Catalogue standing	Score
Arena Elo — Coding	unassessed	1370
Arena Elo — Math	unassessed	1400
Arena Elo — Overall (Text)	unassessed	1360
Artificial Analysis Intelligence Index	unassessed	84
Artificial Analysis Output Speed	unassessed	40
BIG-Bench Hard	unassessed	39.8
GPQA Diamond	unassessed	36.2
HumanEval	unassessed	62.5
IFEval	unassessed	64
MATH-500	unassessed	72.1
MMLU: Astronomy	unassessed	76.5
MMLU: Biology (subcategory)	unassessed	84.2
MMLU: Business Ethics	unassessed	77.8
MMLU: Chemistry (subcategory)	unassessed	78.5
MMLU: Clinical Knowledge	unassessed	83.5
MMLU: Computer Science (subcategory)	unassessed	85.5
MMLU: Jurisprudence	unassessed	74.5
MMLU: Physics (unresolved key)	unassessed	82.8
MMLU-Pro	unassessed	48.5
MMLU: Professional Accounting	unassessed	68.2
MMLU: Professional Law	unassessed	72.5
MuSR	unassessed	14.1

Data

This card as JSON · Edit on GitHub