Gemma 2 9B — Benchmarks

Benchmark scores for Gemma 2 9B aggregated from public leaderboards (4 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.

No llmrun Composite Score for Gemma 2 9B: the score is only published for a model measured on at least 4 of the boards in the score panel, spanning at least 2 benchmark categories and including a reasoning or coding board — or on 3 boards that narrow its score to a 90% range of at most 15.6 points. Either way the results must come from at least two different benchmarks. Its individual scores below stand on their own — see how the llmrun Composite Score gets calculated.

Gemma 2 9B reasoning benchmarks

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

PIQA
PIQA (Physical Interaction QA) asks a model to choose the more sensible of two ways to accomplish an everyday physical task, testing practical, physical-world commonsense that isn't well captured by purely textual reasoning tests.
83.7%
83.7%
median 80.1%open leader 88.6%
Overall best: GPT 4o Mini (Jul 18, 2024) 88.7%
90th percentile · #5 of 40 open · #9 of 59 all
#5 of 40
#9 of 59
Lab-reported
Epoch AI

Gemma 2 9B math benchmarks

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

GSM8K
GSM8K is a classic set of grade-school math word problems requiring several arithmetic reasoning steps. Long a default sanity check for reasoning, top models now score near the ceiling, so it best separates smaller models.
84.9%
84.9%
median 52.3%open leader 94.5%
85th percentile · #10 of 62 open · #15 of 93 all
#10 of 62
#15 of 93
Lab-reported
Epoch AI

Gemma 2 9B 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.
45.1%
45.1%
median 52.1%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
42nd percentile · #85 of 145 open · #178 of 257 all
#85 of 145
#178 of 257
Independent
TIGER-Lab

Other results (1)· leaderboard-only, not part of the llmrun Score

BoolQ
BoolQ is a reading-comprehension benchmark of naturally occurring yes/no questions, each paired with a short passage that contains the answer. It tests whether a model can extract and reason over information from real text rather than just recall facts.
85.7%
85.7%
median 76.4%open leader 89.4%
Overall best: T5 11B 91.2%
84th percentile · #7 of 38 open · #21 of 76 all
#7 of 38
#21 of 76
Lab-reported
Epoch AI

Scores aggregated from public benchmark sources (Epoch AI: 3 boards, 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.