Model

OLMo 2 13B

OPEN

Allen Institute for AI (Ai2) · OLMo 2 family · released Nov 26, 2024

Flagship of Ai2's fully-open Nov 2024 OLMo 2 launch; Instruct variant beats Qwen 2.5 14B Instruct on average score

ReasoningCodingVisionFunction callingTool useAgentic

632.7

Elo · rank #277

Parameters

13B

Active params

13B (dense)

Context

4K tokens

Architecture

Dense Transformer (Olmo2 arch: 40 layers, QK-norm, reordered post-norm)

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

AGIEvalReasoning54.2%#11

best: OLMo 3-Think 32B · 88.2%

AlpacaEvalHuman preference39.5%#27

best: Tülu 2+DPO 70B · 95.1%

ARC-ChallengeReasoning83.5%#36

best: Llama 3.1 405B · 96.9%

BIG-Bench HardReasoning58.8%#81

best: ERNIE 4.5 300B-A47B · 94.3%

DROPReasoning71.5#43

best: Hunyuan-T1 · 93.1

GSM8KMath87.4%#54

best: Llama 3.1 405B · 96.8%

HellaSwagReasoning86.4%#27

best: Claude 3 Opus · 95.4%

IFEvalReasoning82.6%#85

best: Gemma 4 26B A4B · 98.5%

MATH-500Math39.2%#148

best: GPT-5 · 99.4%

MMLU-ProKnowledge35.1%#159

best: Claude Fable 5 · 91.5%

MMLUKnowledge68.5%#156

best: OpenAI o3 · 92.9%

TriviaQAKnowledge81.9%#11

best: Sarvam-1 (2B) · 90.6%

TruthfulQAKnowledge64.3%#17

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

WinoGrandeReasoning81.5%#27

best: PaLM 2 · 90.9%

Run it locally

VRAM @ Q4

8.4 GB

VRAM @ FP16

27.4 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

OLMo 2 family

Elo progression across releases

OLMo 2 13BNov 2024632.7 OLMo 2 7BNov 2024446.1 OLMo 2 32BMar 2025823.2

Resources

HFallenai/OLMo-2-1124-13B GHallenai/OLMo DOCSallenai.org/olmo

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