Model

Alpaca 7B

OPEN

LLaMA 7B fine-tuned on 52K self-instruct demos for under $600; sparked the open instruction-tuning wave.

ReasoningCodingVisionFunction callingTool useAgentic

-229.7

Elo · rank #427

Parameters

6.7B

Active params

6.7B (dense)

Context

0.512K tokens

Architecture

Dense Transformer (LLaMA 7B architecture, fine-tuned)

License

CC-BY-NC-4.0 (data/weights, non-commercial); Apache-2.0 (training code)

Languages

API price (in/out)

No hosted API

Modalities

text

Benchmark results

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

AlpacaEvalHuman preference5.9%#37

best: Tülu 2+DPO 70B · 95.1%

BIG-Bench HardReasoning33.3%#122

best: ERNIE 4.5 300B-A47B · 94.3%

HumanEvalCoding10.3%#183

best: Claude Opus 4.5 · 99.4%

MMLUKnowledge41.6%#255

best: OpenAI o3 · 92.9%

Run it locally

VRAM @ Q4

6 GB

VRAM @ FP16

13 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

Alpaca family

Elo progression across releases

Resources

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