Compare · ModelsLive · 2 picked · head to head
Gemma 3 27B (free) vs o3 Mini
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
o3 Mini wins on 5/6 benchmarks
o3 Mini wins 5 of 6 shared benchmarks. Leads in coding · knowledge · math.
Category leads
coding·o3 Miniknowledge·o3 Minimath·o3 Mini
Hype vs Reality
Attention vs performance
Gemma 3 27B (free)
#172 by perf·#7 by attention
o3 Mini
#239 by perf·no signal
Vendor risk
Who is behind the model
Google DeepMind
$4.20T·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
6 benchmarks · 2 models
Gemma 3 27B (free)o3 Mini
Aider polyglot
o3 Mini leads by +55.5
Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework.
Gemma 3 27B (free)
4.9
o3 Mini
60.4
Fiction.LiveBench
o3 Mini leads by +16.7
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
Gemma 3 27B (free)
33.3
o3 Mini
50.0
GPQA diamond
o3 Mini leads by +37.5
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Gemma 3 27B (free)
31.8
o3 Mini
69.4
Lech Mazur Writing
Gemma 3 27B (free) leads by +18.2
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
Gemma 3 27B (free)
79.9
o3 Mini
61.7
MATH level 5
o3 Mini leads by +22.4
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
Gemma 3 27B (free)
74.0
o3 Mini
96.5
OTIS Mock AIME 2024-2025
o3 Mini leads by +57.3
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Gemma 3 27B (free)
19.6
o3 Mini
76.9
Full benchmark table
| Benchmark | Gemma 3 27B (free) | o3 Mini |
|---|---|---|
Aider polyglot Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework. | 4.9 | 60.4 |
Fiction.LiveBench Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination. | 33.3 | 50.0 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 31.8 | 69.4 |
Lech Mazur Writing Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication. | 79.9 | 61.7 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 74.0 | 96.5 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 19.6 | 76.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.00 | $0.00 | 131K tokens (~66 books) | — | |
| $1.10 | $4.40 | 200K tokens (~100 books) | $19.25 |
People also compared