Request a pilot
‹ PublicAI Index

The LLM benchmark aggregator.

Gemma 3 27B It

Google · 27B · open weights

Strongest in Factual grounding (#36 of 101), weakest in Agents (#255 of 268). Above par in 6 of 16 scopes. Among the models it meets almost everywhere, it finishes behind Claude Sonnet 5 and Inkling and ahead of Mistral Large and Nova Micro.

§ 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.

Safety51.6#142/337
Factual grounding56.5#36/101
Jailbreak resistance56.8#85/272
Toxicity avoidance55.3#99/272
Fairness46.7#157/300
Harm refusal49.8#160/300
Secure code48.8#169/274
Coding41.5#146/165
Agentic coding39.8#144/157
Human preference53.6#151/342
Human preference53.6#151/342
Core abilities44#160/204
General intelligence41.4#163/204
Agents38.7#255/268
Tool use43.9#49/81
Knowledge work30.5#175/178

§ 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 Gemma 3 27B It, left for the other.

Magistral Mediumwon 10 · lost 0
Gemma 3 4B Itwon 11 · lost 2
Llama 3.2 3B Instructwon 8 · lost 2
QwQ 32Bwon 8 · lost 2
Command Awon 12 · lost 3
Llama 3.1 8B Instructwon 8 · lost 3
Grok Build 0.1won 8 · lost 4
Grok 4.3won 9 · lost 5
Claude 3 Haikuwon 7 · lost 4
Nova Prowon 7 · lost 4
Mistral Smallwon 8 · lost 5
Grok 4.1 Fastwon 8 · lost 5
Ling 3.0 Flashwon 6 · lost 4
Llama 3.2 1B Instructwon 6 · lost 4
Mimo V2.6 Flashwon 6 · lost 4
Llama 4 Scout Instructwon 8 · lost 6
Nova Microwon 6 · lost 5
Mistral Largewon 6 · lost 5
Ling 3.0 Flash VLwon 5 · lost 5
Llama 3 70B Instructwon 5 · lost 5
Qwen3 30B A3B Instructwon 5 · lost 5
GLM-4.5won 5 · lost 5
Kimi K2.7 Codewon 6 · lost 6
Gemma 3 12B Itwon 6 · lost 6
GPT-4o Miniwon 7 · lost 7
GPT-4.1 Nanowon 7 · lost 8
Gemma 4 E4Bwon 5 · lost 6
Gemma 4 12Bwon 5 · lost 6
Grok 3won 5 · lost 6
Gemma 4 26B A4Bwon 5 · lost 6
GPT OSS 20Bwon 5 · lost 7
Gemini 2.0 Flashwon 5 · lost 7
Qwen3 8Bwon 4 · lost 6
GPT-4won 4 · lost 6
Mistral Large 3won 4 · lost 6
Gemini 3.1 Flash Litewon 4 · lost 6
DeepSeek V3won 6 · lost 9
Kimi K2.5won 6 · lost 9
Kimi K2.6won 6 · lost 9
GLM-4.6won 6 · lost 10
Phi 4won 4 · lost 7
GPT-4owon 4 · lost 7
DeepSeek V4 Flashwon 5 · lost 9
Gemini 3.6 Flashwon 5 · lost 9
GPT OSS 120Bwon 5 · lost 10
Gemini 2.5 Flashwon 5 · lost 10
DeepSeek R1won 5 · lost 10
Gemini 2.5 Prowon 5 · lost 10
DeepSeek V4 Prowon 5 · lost 10
GLM-5.2won 5 · lost 10
Gemma 4 31Bwon 5 · lost 11
GPT-5 Nanowon 4 · lost 9
Kimi K2 Thinkingwon 3 · lost 7
Qwen Maxwon 3 · lost 7
O1 Miniwon 3 · lost 7
Grok 4won 4 · lost 10
Qwen3.7 Maxwon 4 · lost 10
Gemini 3.5 Flashwon 4 · lost 10
MiniMax M3won 4 · lost 11
Claude 3.5 Haikuwon 3 · lost 9
O3 Miniwon 3 · lost 9
DeepSeek V3.2won 4 · lost 12
Qwen3.5 122B A10Bwon 3 · lost 10
GPT-5.1won 3 · lost 10
Qwen3 235B A22B Instructwon 3 · lost 11
Mimo V2.5 Prowon 3 · lost 11
GLM-5.3 Flashwon 3 · lost 11
Gemini 3.8 Flashwon 3 · lost 11
MiniMax M2won 2 · lost 8
GPT-5.4 Nanowon 2 · lost 8
GPT-5.4 Miniwon 2 · lost 8
Claude Opus 4.5won 2 · lost 8
GPT-6 Astrawon 2 · lost 8
Gemini 3.1 Prowon 2 · lost 8
Kimi K2won 3 · lost 12
GPT-5.5won 3 · lost 12
GPT-5.6 Solwon 3 · lost 12
Qwen3 32Bwon 2 · lost 9
Qwen3.5 35B A3Bwon 2 · lost 9
Claude Opus 4.6won 2 · lost 9
GLM-5.1won 2 · lost 10
Gemini 3 Prowon 2 · lost 10
GPT-6 Lunawon 2 · lost 12
Grok 4.6won 2 · lost 12
Grok 4.5won 2 · lost 12
Qwen3.8 Maxwon 2 · lost 12
GPT-4.1 Miniwon 2 · lost 13
O4 Miniwon 2 · lost 13
GPT-4.1won 2 · lost 13
GPT-5won 2 · lost 13
GPT-5.2won 2 · lost 13
GPT-5.6 Lunawon 2 · lost 13
GPT-6 Solwon 2 · lost 13
GPT-5.6 Terrawon 2 · lost 13
GPT-5.4won 2 · lost 13
GPT-5 Miniwon 2 · lost 14
Claude Haiku 4.5won 2 · lost 14
Claude Sonnet 4.5won 2 · lost 14
Claude 3.7 Sonnetwon 1 · lost 10
Qwen3.5 397B A17Bwon 1 · lost 11
Hy3won 1 · lost 11
Claude Opus 4won 1 · lost 12
Claude Sonnet 4.6won 1 · lost 12
O3won 1 · lost 13
Mimo V2.6 Prowon 1 · lost 13
Claude Opus 5won 1 · lost 13
Muse Spark 1.1won 1 · lost 13
Claude Opus 5.5won 1 · lost 13
Claude Sonnet 4won 1 · lost 14
Claude Opus 4.7won 1 · lost 14
Claude 3.5 Sonnetwon 0 · lost 10
Gemini 2.5 Flash Litewon 0 · lost 10
O1won 0 · lost 10
MiniMax M2.7won 0 · lost 12
Grok 4.7won 0 · lost 14
Claude Opus 4.8won 0 · lost 14
Inklingwon 0 · lost 15
Claude Sonnet 5won 0 · lost 15

§ 3 · Sources

Where the numbers come from

7 publications, 12 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26
LMArena Text 1365
GDPval-AA ↗1 measureread 2026-09-26
GDPval-AA -424
Aider polyglot ↗1 measureread 2026-09-26
Aider polyglot 4.9%
BFCL ↗1 measureread 2026-09-26
BFCL v4 29.47%
Kagi LLM Benchmark ↗1 measureread 2026-09-26
Kagi LLM Benchmark 35.1%
Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-09-26
Enkrypt · Jailbreak risk 7%Enkrypt · Harmful content risk 43.9%Enkrypt · CBRN risk 9.7%Enkrypt · Toxicity risk 2.6%Enkrypt · Bias risk 83%Enkrypt · Insecure code risk 36%
Vectara · Factual consistency 92.6%

Badge

PublicAI Index badge for Gemma 3 27B It[![PublicAI Index](https://publicai.io/model-index/badge?model=gemma-3-27b-it)](https://publicai.io/model-index/m/gemma-3-27b-it)