Model

Llama 2 7B Chat

OPEN

Meta · Llama 2 family · released Jul 18, 2023

RLHF dialogue-tuned Llama 2 7B; 4K context; MMLU 45.3, far behind later Llama generations.

ReasoningCodingVisionFunction callingTool useAgentic

-10.6

Elo · rank #406

Parameters

Active params

7B (dense)

Context

4K tokens

Architecture

Dense transformer (decoder-only), RLHF-tuned

License

Llama 2 Community License

Languages

—

API price (in/out)

No hosted API

Modalities

text

Benchmark results

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

ARC-ChallengeReasoning52.9%#116

best: Llama 3.1 405B · 96.9%

Arena-HardHuman preference4.6%#54

best: Qwen3-235B-A22B (Non-Thinking) · 96.1%

Arena EloHuman preference1037#52

best: Claude Fable 5 · 1505

DROPReasoning13.09#64

best: Hunyuan-T1 · 93.1

GSM8KMath7.3%#181

best: Llama 3.1 405B · 96.8%

HellaSwagReasoning78.6%#90

best: Claude 3 Opus · 95.4%

MMLU (EU-21 languages)Knowledge37.0%#14

best: Llama 3.1 70B · 77.1%

MMLUKnowledge45.8%#243

best: OpenAI o3 · 92.9%

MT-BenchHuman preference6.27#64

best: Hunyuan-Large (A52B) · 9.4

RakudaHuman preference2.08#7

best: PLaMo-100B · 9.73

TruthfulQAKnowledge45.6%#64

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

WinoGrandeReasoning71.7%#90

best: PaLM 2 · 90.9%

Run it locally

VRAM @ Q4

4.1 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 · AWQ · MLX

Llama 2 family

Elo progression across releases

Llama 2 13BJul 202369.7 Llama 2 13B ChatJul 2023105.1 Llama 2 70BJul 2023365.0 Llama 2 70B ChatJul 2023363.8 Llama 2 7BJul 2023-121.8 Llama 2 7B ChatJul 2023-10.6

Resources

HFmeta-llama/Llama-2-7b-chat-hf GHfacebookresearch/llama DOCSllama.com/llama2 APIAvailable via Meta API

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