Model
Model explorer

OpenLLaMA 7B

OPEN
OpenLM Research (Berkeley) · OpenLLaMA family · released May 3, 2023

Apache-2.0 from-scratch reproduction of Meta LLaMA on RedPajama (3B/7B/13B, plus v2); a permissive drop-in base free of LLaMA's license.

ReasoningCodingVisionFunction callingTool useAgentic
-322.0
Elo · rank #434
Parameters
6.7B
Active params
6.7B (dense)
Context
2K tokens
Architecture
Dense LLaMA-architecture transformer decoder, from-scratch reproduction trained on RedPajama
License
Apache 2.0
Languages
—
API price (in/out)
No hosted API
Modalities
text
Benchmark results
Bar shows position within the tracked field; marker = field best
ARC-ChallengeReasoning38.0%#153
best: Llama 3.1 405B · 96.9%
ARC-EasyReasoning68.0%#53
best: Phi-3-medium (14B) · 97.7%
DROPReasoning5.5#72
best: Hunyuan-T1 · 93.1
GSM8KMath1.6%#192
best: Llama 3.1 405B · 96.8%
HellaSwagReasoning72.0%#119
best: Claude 3 Opus · 95.4%
MMLUKnowledge30.6%#269
best: OpenAI o3 · 92.9%
OpenBookQAReasoning40.0%#44
best: Claude 1 · 90.8%
PIQAReasoning77.0%#62
best: GPT-4o mini · 93.1%
TruthfulQAKnowledge35.0%#96
best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%
WinoGrandeReasoning67.0%#111
best: PaLM 2 · 90.9%
Run it locally
VRAM @ Q4
4.5 GB
VRAM @ FP16
13.5 GB
Fits on (Q4)
RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB
Throughput data unavailable.
Quantizations
GGUF Q4 · GPTQ
OpenLLaMA family
Elo progression across releases
OpenLLaMA 7BMay 2023-322.0
Resources
HFopenlm-research/open_llama_7bGHopenlm-research/open_llama
API price weights · each benchmark row carries its own source badge (see methodology)