Gemma 3 27B IT — Benchmarks

Benchmark scores for Gemma 3 27B IT aggregated from public leaderboards (11 boards), with how it ranks among open models and how it compares with its nearest neighbours. See hardware requirements for what you need to run it.

llmrun Composite Score →

79.4

90% range 75.6–83.3 · from 9 boards

Coding69.790% range 54–85
AgentsNot enough data yet
MathNot enough data yet
Science80.790% range 72–89
Reasoning73.690% range 62–85

Gemma 3 27B IT reasoning benchmarks

Multi-step logic and problem-solving — can the model think through a hard problem rather than recall an answer.

GPQA Diamond
GPQA Diamond is a set of extremely hard, graduate-level science questions (physics, chemistry, biology) written by domain experts and filtered so that skilled non-experts with web access still fail. It measures genuine reasoning rather than memorization.
47.7%
47.7%
median 54.0%open leader 93.1%
Overall best: Gemini 4 Argon (high) 96.4%
42nd percentile · #57 of 98 open · #234 of 300 all
#57 of 98
#234 of 300
Independent
Epoch AI · Mar 2025
CritPt
CritPt (Critical Points) tests graduate-level physics problem-solving on research-style questions spanning mechanics, electromagnetism, quantum mechanics, and more. It was created by physicists to probe whether a model can reason through multi-step physics rather than pattern-match textbook answers.
0.0%
0.0%
median 1.1%open leader 26.6%
Overall best: GPT 5.6 Sol Max 32.3%
15th percentile · #50 of 66 open · #187 of 219 all
#50 of 66
#187 of 219
Lab-reported
Epoch AI · Mar 2025
DTBench
DTBench is a set of handcrafted, expert-validated multiple-choice questions on decision theory — Newcomb-like problems and related puzzles — built by the Conceptual Reasoning Index project with Anthropic. It tests whether a model can follow an abstract argument rather than recall a fact.
52.5%
52.5%
median 67.7%open leader 93.9%
Overall best: Claude Opus 5.5 Max 98.9%
16th percentile · #57 of 68 open · #190 of 214 all
#57 of 68
#190 of 214
Lab-reported
Epoch AI · Mar 2025
Fiction.LiveBench
Fiction.LiveBench is a long-context comprehension test: models answer questions that require tracking characters, events and relationships across a long story. llmrun shows the 16k-token score, the column Epoch AI reports as its headline.
33.3%
33.3%
median 61.1%open leader 86.1%
Overall best: GPT 5 (Aug 07, 2025, medium) 97.2%
6th percentile · #23 of 25 open · #55 of 58 all
#23 of 25
#55 of 58
Independent
Epoch AI · Mar 2025
LMCA
LMCA (Language Model Conceptual Argumentation) scores models on expert-rated arguments across philosophy, decision theory, and AI risk — domains where there is no empirical feedback loop to check an answer against. It is part of the Conceptual Reasoning Index, built with Anthropic by researchers Emery Cooper and Caspar Oesterheld.
12.3%
12.3%
median 24.7%open leader 55.5%
Overall best: Claude Opus 5.5 Max 68.2%
17th percentile · #45 of 54 open · #159 of 176 all
#45 of 54
#159 of 176
Lab-reported
Epoch AI · Mar 2025

Gemma 3 27B IT coding benchmarks

Writing, editing, and fixing real code — measured by whether the resulting program actually runs and passes tests.

Aider Polyglot
The Aider Polyglot benchmark measures real-world coding across several programming languages: the model edits code to solve Exercism exercises, and is scored on whether the final solution actually runs and passes the tests.
4.9%
4.9%
median 45.1%open leader 74.2%
Overall best: GPT 5 (Aug 07, 2025, high) 88.0%
lowest · #18 of 18 open · #68 of 69 all
#18 of 18
#68 of 69
Independent
Epoch AI · Mar 2025
SciCode
SciCode asks a model to write real scientific-computing code — implementing multi-step numerical methods drawn from physics, chemistry, biology, and math research — and checks the output against reference solutions. It was built by a team of scientists and ML researchers to test coding ability on genuine research problems, not typical software tasks.
23.3%
23.3%
median 38.7%open leader 60.9%
Overall best: Claude Opus 5.5 Max 66.9%
8th percentile · #61 of 66 open · #202 of 207 all
#61 of 66
#202 of 207
Lab-reported
Epoch AI · Mar 2025

Gemma 3 27B IT math benchmarks

Symbolic and quantitative problem-solving, from competition math to multi-step calculation.

AIME 2024/2025
AIME (American Invitational Mathematics Examination) is a prestigious high-school competition of hard, integer-answer problems. It's a widely-cited yardstick for multi-step mathematical reasoning, here on the 2024–2025 papers.
22.5%
22.5%
median 52.8%open leader 98.6%
Overall best: Claude Fable 5 (high) 100.0%
40th percentile · #52 of 86 open · #222 of 283 all
#52 of 86
#222 of 283
Independent
Epoch AI · Mar 2025
MATH Level 5
MATH Level 5 covers the hardest tier of competition-mathematics problems, testing multi-step symbolic and quantitative reasoning.
74.0%
74.0%
median 41.6%open leader 96.6%
Overall best: GPT 5 (Aug 07, 2025, high) 98.1%
87th percentile · #6 of 39 open · #41 of 107 all
#6 of 39
#41 of 107
Independent
Epoch AI · Mar 2025

Gemma 3 27B IT knowledge benchmarks

Breadth and depth of factual knowledge across many subjects, paired with the reasoning to apply it.

MMLU-Pro
MMLU-Pro is a harder, cleaned-up successor to MMLU with ten answer choices and more reasoning-heavy questions across 14 subjects, measuring broad knowledge and reasoning together.
67.5%
67.5%
median 52.1%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
71st percentile · #43 of 145 open · #109 of 257 all
#43 of 145
#109 of 257
Independent
TIGER-Lab

Gemma 3 27B IT instruction following benchmarks

How precisely a model obeys explicit, verifiable instructions — format, length, and content constraints — and, more broadly, how much people prefer its responses head-to-head in blind human voting.

LMArena Text
Arena (arena.ai, formerly LMArena / Chatbot Arena) ranks models by millions of blind, head-to-head human votes on real prompts. The Arena Score shown here is a Bradley-Terry rating computed with Arena's default style control, so higher means people preferred that model's answers more often — it's a relative rating, not a percentage, so compare it only to other models on this board.
1365.5
1365.5
median 1321.5open leader 1487.1
Overall best: Gemini 4 Argon (High) 1523.9
65th percentile · #66 of 187 open · #191 of 405 all
#66 of 187
#191 of 405
Arena
Arena · Oct 2026

Score vs VRAM in the 24 GB tier

Gemma 3 27B IT
Other models
Best Score for its VRAM
x: VRAM (GB) · y: llmrun Score

#25 of 35 that fit 24 GB

Each dot is a model at the best quant that fits a 24 GB GPU. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.

Full 24 GB ranking →

Gemma 3 27B IT vs its neighbours

Same boards for the nearest open models per llmrun Score and the best model that fits 24 GB. The highest score in each row is marked.

Board	Gemma 3 27B ITthis model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	Qwen2.5 Coder 32B Instructnearest by ScoreQ4_K_S · 23 GB	Qwen2.5 32B Instructnearest by ScoreQ4_K_S · 23 GB	Mistral Small 3.1 24B Instruct 2503nearest by ScoreQ4_K_M · 18 GB
llmrun Score	79.4	122.1	81.9	80.4	79.0
GPQA Diamond	47.7%	—	—	46.1%	47.5%
CritPt	0.0%	5.4%	—	—	0.0%
DTBench	52.5%	88.0%	—	—	58.6%
LMCA	12.3%	41.4%	—	—	—
Aider Polyglot	4.9%	—	16.4%	—	—
SciCode	23.3%	46.6%	—	—	27.8%
AIME 2024/2025	22.5%	—	—	7.4%	5.8%
MATH Level 5	74.0%	—	—	56.1%	46.8%
MMLU-Pro	67.5%	—	—	69.2%	—
LMArena Text	1365.5	1437.8	1270.6	—	1302.7
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 9 boards, Arena: 1 board, TIGER-Lab: 1 board) — each linked from the benchmark pages. llmrun does not run these benchmarks. Percentiles and ranks are among open models in the llmrun catalogue; the median and leader marks refer to the same field.