The LLM benchmark aggregator.

Snowflake Arctic Instruct

Other

Strongest in Factual grounding (#7 of 104), weakest in Human preference (#300 of 354). Above par in 2 of 4 scopes. Among the models it meets almost everywhere, it finishes behind DeepSeek V4.1 Flash and GPT-5.4 nano and ahead of Llama 2 7B Chat and Granite 3.1 2B Instruct.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Safety54.5−6.9#61/3271/3

Factual grounding64.3−6.4#7/1041/1

Human preference35.5−33.3#300/3541/1

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Snowflake Arctic Instruct, left for the other.

OLMo 7B Instructwon 3 · lost 0

Gemma 2B ITwon 3 · lost 0

Llama 3.2 1B Instructwon 3 · lost 0

SmolLM2 1.7B Instructwon 3 · lost 0

Phi-3 Mini 128K Instructwon 3 · lost 0

Zephyr 7B Betawon 3 · lost 0

Gemma 7B ITwon 3 · lost 0

Llama 2 13B Chatwon 3 · lost 0

Phi-3 Mini 4K Instructwon 3 · lost 0

Mistral 7B Instruct v0.2won 3 · lost 0

Granite 3.0 2B Instructwon 3 · lost 0

Nous Hermes 2 Mixtral 8x7B DPOwon 3 · lost 0

Llama 3.2 3B Instructwon 3 · lost 0

Llama 2 70B Chatwon 3 · lost 0

Phi-3 Small 8K Instructwon 3 · lost 0

Starling LM 7B Betawon 3 · lost 0

Granite 3.1 2B Instructwon 3 · lost 0

Llama 2 7B Chatwon 2 · lost 1

Aya Expanse 8Bwon 2 · lost 2

Ministral 8Bwon 2 · lost 2

Amazon Nova Microwon 2 · lost 2

Amazon Nova Litewon 2 · lost 2

Aya Expanse 32Bwon 2 · lost 2

Mistral Small 3 24B Instructwon 2 · lost 2

Command R+won 2 · lost 2

Amazon Nova Prowon 2 · lost 2

Gemma 3 4B ITwon 2 · lost 2

Mistral Large 2.1won 2 · lost 2

NVIDIA Nemotron 3 Nano 30B A3Bwon 2 · lost 2

Llama 4 Scout Instructwon 2 · lost 2

Llama 4 Maverick Instructwon 2 · lost 2

Amazon Nova 2 Litewon 2 · lost 2

Gemma 3 12B ITwon 2 · lost 2

Mercury 2won 2 · lost 2

Qwen3 32Bwon 2 · lost 2

GPT OSS 120Bwon 2 · lost 2

Command Awon 2 · lost 2

GLM-4.7 Flashwon 2 · lost 2

Gemma 3 27B ITwon 2 · lost 2

Trinity Largewon 2 · lost 2

GLM-4.5 Airwon 2 · lost 2

MiniMax M2.1won 2 · lost 2

MiniMax M2.5won 2 · lost 2

O4 Miniwon 2 · lost 2

Qwen3.5 35B A3Bwon 2 · lost 2

DeepSeek V3won 2 · lost 2

Qwen3.5 Flashwon 2 · lost 2

Grok 4 Fastwon 2 · lost 2

Mistral Medium 3.1won 2 · lost 2

Qwen3.5 27Bwon 2 · lost 2

Gemini 2.5 Flashwon 2 · lost 2

Grok 3won 2 · lost 2

Mistral Large 3won 2 · lost 2

DeepSeek V3.1won 2 · lost 2

Qwen3.5 122B A10Bwon 2 · lost 2

Kimi K2won 2 · lost 2

DeepSeek R1won 2 · lost 2

Qwen3 235B A22B Instructwon 2 · lost 2

DeepSeek V3.2won 2 · lost 2

GLM-4.6won 2 · lost 2

Grok 4.1 Fastwon 2 · lost 2

Gemini 3.1 Flash-Litewon 2 · lost 2

Gemma 4 26B A4Bwon 2 · lost 2

GLM-4.7won 2 · lost 2

Gemini 2.5 Prowon 2 · lost 2

GPT-5.4 miniwon 2 · lost 2

Kimi K2.5won 2 · lost 2

Claude Opus 4.1won 2 · lost 2

Gemma 4 31Bwon 2 · lost 2

GPT-6 Solwon 2 · lost 2

GLM-5won 2 · lost 2

Kimi K2.6won 2 · lost 2

DeepSeek V4 Prowon 2 · lost 2

Claude Sonnet 4.6won 2 · lost 2

Gemini 3 Flashwon 2 · lost 2

GPT-6 Astrawon 2 · lost 2

Claude Opus 4.5won 2 · lost 2

GPT-5.6 Solwon 2 · lost 2

Gemini 3 Prowon 2 · lost 2

Gemini 3.1 Prowon 2 · lost 2

Claude Opus 4.6won 2 · lost 2

Gemini 4 Argonwon 2 · lost 2

Granite 3.0 8B Instructwon 1 · lost 2

DeepSeek LLM 67B Chatwon 1 · lost 2

Qwen1.5 14B Chatwon 1 · lost 2

DBRX Instructwon 1 · lost 2

Mixtral 8x7B Instructwon 1 · lost 2

Phi-3 Medium 4K Instructwon 1 · lost 2

Gemma 2 2B ITwon 1 · lost 2

GPT-3.5 Turbowon 1 · lost 2

Granite 3.1 8B Instructwon 1 · lost 2

Llama 3.1 8B Instructwon 1 · lost 2

Llama 3.1 Tulu 3 8Bwon 1 · lost 2

Llama 3 8B Instructwon 1 · lost 2

Mixtral 8x22B Instructwon 1 · lost 2

Qwen1.5 72B Chatwon 1 · lost 2

Jamba 1.5 Miniwon 1 · lost 2

Command Rwon 1 · lost 2

OLMo 2 32B Instructwon 1 · lost 2

Claude 3 Haikuwon 1 · lost 2

Gemma 2 9B ITwon 1 · lost 2

GPT-4won 1 · lost 2

Llama 3 70B Instructwon 1 · lost 2

Jamba 1.5 Largewon 1 · lost 2

Llama 3.1 70B Instructwon 1 · lost 2

Llama 3.1 Nemotron 70B Instructwon 1 · lost 2

Magistral Mediumwon 1 · lost 2

Gemini 1.5 Flashwon 1 · lost 2

GPT-4o miniwon 1 · lost 2

Qwen Maxwon 1 · lost 2

GPT OSS 20Bwon 1 · lost 2

GPT-4.1 nanowon 1 · lost 2

Llama 3.1 405B Instructwon 1 · lost 2

QwQ 32Bwon 1 · lost 2

Gemini 1.5 Prowon 1 · lost 2

Gemini 2.0 Flash-Litewon 1 · lost 2

NVIDIA Nemotron 3 Super 120B A12Bwon 1 · lost 2

Gemini 2.0 Flashwon 1 · lost 2

Grok 3 Miniwon 1 · lost 2

Qwen3 30B A3B Instructwon 1 · lost 2

GPT-4.1 miniwon 1 · lost 2

GLM-4.5won 1 · lost 2

Grok 4won 1 · lost 2

Qwen3 VL 235B A22B Instructwon 1 · lost 2

Mistral Large 4won 1 · lost 2

NVIDIA Nemotron 3 Ultra 550B A55Bwon 1 · lost 2

MiniMax M3won 1 · lost 2

DeepSeek V4 Flashwon 1 · lost 2

GPT-6 Lunawon 1 · lost 2

Grok 4.3won 1 · lost 2

MiMo V2.6 Flashwon 1 · lost 2

GPT-5.6 Lunawon 1 · lost 2

GLM-5.1won 1 · lost 2

MiMo V2.5 Prowon 1 · lost 2

Qwen3.7 Maxwon 1 · lost 2

GLM-5.3 Flashwon 1 · lost 2

GLM-5.2won 1 · lost 2

Gemini 3.5 Flashwon 1 · lost 2

MiMo V2.6 Prowon 1 · lost 2

Gemini 3.6 Flashwon 1 · lost 2

Qwen3.8 Maxwon 1 · lost 2

Gemini 3.8 Flashwon 1 · lost 2

Granite 4.0 H Smallwon 1 · lost 3

Llama 3.3 70B Instructwon 1 · lost 3

GPT-4owon 1 · lost 3

GPT-5 nanowon 1 · lost 3

Qwen3 Next 80B A3B Thinkingwon 1 · lost 3

Gemini 2.5 Flash-Litewon 1 · lost 3

GPT-5 miniwon 1 · lost 3

Claude Sonnet 4won 1 · lost 3

Claude Haiku 4.5won 1 · lost 3

GPT-4.1won 1 · lost 3

MiniMax M2.7won 1 · lost 3

Claude Opus 4won 1 · lost 3

GPT-5won 1 · lost 3

GPT-5.2won 1 · lost 3

GPT-5.1won 1 · lost 3

Claude Sonnet 4.5won 1 · lost 3

GPT-5.4won 1 · lost 3

GPT-5.5won 1 · lost 3

Claude Opus 4.7won 1 · lost 3

Claude Opus 5.5won 1 · lost 3

Qwen2 72B Instructwon 0 · lost 3

Claude 3 Sonnetwon 0 · lost 3

Gemma 2 27B ITwon 0 · lost 3

Qwen2.5 72B Instructwon 0 · lost 3

Claude 3 Opuswon 0 · lost 3

Claude 3.5 Haikuwon 0 · lost 3

GPT-4 Turbowon 0 · lost 3

O1 Miniwon 0 · lost 3

MiniMax M2won 0 · lost 3

O3 Miniwon 0 · lost 3

Claude 3.5 Sonnetwon 0 · lost 3

Claude 3.7 Sonnetwon 0 · lost 3

O1won 0 · lost 3

O3won 0 · lost 3

Qwen3.5 397B A17Bwon 0 · lost 3

Grok 4.7won 0 · lost 3

Inklingwon 0 · lost 3

GPT-4.5won 0 · lost 3

Hy3won 0 · lost 3

Grok 4.6won 0 · lost 3

Claude Sonnet 5won 0 · lost 3

Grok 4.5won 0 · lost 3

GPT-5.6 Terrawon 0 · lost 3

Claude Opus 4.8won 0 · lost 3

Claude Opus 5won 0 · lost 3

Muse Spark 1.1won 0 · lost 3

Phi 4won 0 · lost 4

GPT-5.4 nanowon 0 · lost 4

DeepSeek V4.1 Flashwon 0 · lost 4

§ 3 · Sources

Where the numbers come from

2 publications, 2 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-11-12

LMArena Text 1180

Vectara Hallucination Leaderboard ↗1 measureread 2026-11-12

Vectara · Factual consistency 95.7%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=snowflake-arctic-instruct)](https://publicai.io/model-index/m/snowflake-arctic-instruct)