Model

LLaMA 7B

OPEN

Meta · LLaMA family · released Feb 24, 2023

Smallest first-gen LLaMA, trained on roughly 1T tokens for efficient research use under a gated non-commercial license.

ReasoningCodingVisionFunction callingTool useAgentic

-180.9

Elo · rank #421

Parameters

6.7B

Active params

6.7B (dense)

Context

2K tokens

Architecture

Dense decoder-only Transformer

License

LLaMA Non-Commercial Research License (gated application access)

Languages

—

API price (in/out)

No hosted API

Modalities

text

Benchmark results

Bar shows position within the tracked field; marker = field best

AGIEvalReasoning23.9%#41

best: OLMo 3-Think 32B · 88.2%

ARC-ChallengeReasoning47.6%#130

best: Llama 3.1 405B · 96.9%

ARC-EasyReasoning72.8%#42

best: Phi-3-medium (14B) · 97.7%

BIG-Bench HardReasoning30.3%#132

best: ERNIE 4.5 300B-A47B · 94.3%

GSM8KMath11.0%#177

best: Llama 3.1 405B · 96.8%

HellaSwagReasoning76.1%#108

best: Claude 3 Opus · 95.4%

APIBench (HuggingFace)Agents0.0%#5

best: Gorilla 7B · 71.7%

HumanEvalCoding10.5%#182

best: Claude Opus 4.5 · 99.4%

MATH-500Math2.9%#202

best: GPT-5 · 99.4%

MBPPCoding17.7%#107

best: Llama-3.3-Nemotron-Super-49B v1 (Reasoning On) · 91.3%

MMLUKnowledge35.1%#263

best: OpenAI o3 · 92.9%

OpenBookQAReasoning57.2%#17

best: Claude 1 · 90.8%

PIQAReasoning79.8%#47

best: GPT-4o mini · 93.1%

RACE-HReasoning46.9%#16

best: Claude 3 Opus · 92.9%

Social IQaReasoning48.9%#29

best: Apple DCLM-Baseline 7B · 82.9%

APIBench (TensorHub)Agents0.0%#5

best: Gorilla 7B · 83.8%

APIBench (TorchHub)Agents0.0%#5

best: Gorilla 7B · 59.1%

TriviaQAKnowledge50.0%#43

best: Sarvam-1 (2B) · 90.6%

TruthfulQAKnowledge27.4%#106

best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%

WinoGrandeReasoning70.1%#103

best: PaLM 2 · 90.9%

Run it locally

VRAM @ Q4

6.6 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

No verifiably-sourced LLaMA-1-7B-specific RTX 4090 llama.cpp benchmark was found; do not conflate with unrelated Llama-3/Mistral 7B figures circulating online.

Quantizations

GGUF · GPTQ · AWQ · GGML

LLaMA family

Elo progression across releases

LLaMA 13BFeb 20230.6 LLaMA 33BFeb 2023184.9 LLaMA 65BFeb 2023240.9 LLaMA 7BFeb 2023-180.9

Resources

HFhuggyllama/llama-7b GHmeta-llama/llama DOCSai.meta.com/blog/large-language-model-llama-meta-ai APIAvailable via Meta API

API price weights · each benchmark row carries its own source badge (see methodology)