gemma-3-12b-it: benchmark scores, price and ranking
gemma-3-12b-it from Google DeepMind ranks #739 of 743 LLMs on the LLMs Tiger aggregate benchmark score (39.1), based on 7 benchmark results captured up to 2026-10-13.
- Organisation
- Google DeepMind
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2025-03-12 (source: epoch)
- Cheapest API price
- $0.050 in / $0.100 out per million tokens (via novita, observed 2026-10-13T20:47Z)
- Aggregate rank
- #739 of 743 (score 39.1)
Relative to other models, gemma-3-12b-it is strongest on OTIS Mock AIME (#132 of 189) and weakest on LMCA (#107 of 108).
Across the 7 benchmarks where at least ten models are scored, gemma-3-12b-it is in the top 10% on 0 and in the top half on 0.
Its cheapest listed input price, $0.050 per million tokens, is lower than that of 90% of the 427 models with a listed price.
gemma-3-12b-it benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| OTIS Mock AIME | 16.7% | #132 of 189 | Epoch AI Benchmarking Hub | 2026-10-13 |
| CritPt | 0.0% | #91 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GPQA Diamond | 39.5% | #224 of 275 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Chess Puzzles | 0.0% | #101 of 122 | Epoch AI Benchmarking Hub | 2026-10-13 |
| DTBench | 48.8% | #134 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 17.4% | #103 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMCA | 4.5% | #107 of 108 | Epoch AI Benchmarking Hub | 2026-10-13 |
About gemma-3-12b-it
gemma-3-12b-it is an open-weight model that can run on your own hardware from Google DeepMind. It was released on 2025-03-12, according to Epoch AI. The cheapest API listing we track charges $0.050 per million input tokens and $0.100 per million output tokens, through novita. We hold 7 benchmark results for it across 4 areas, and none is in the top tenth of its board.
Where gemma-3-12b-it stands
On our aggregate score, gemma-3-12b-it is #739 of 743 ranked models, in the lower half. Its score is 39.1; the scale is centred on 50, so it sits 10.9 points below the typical model. That beats 1% of the ranked field. Among the 41 ranked Google DeepMind models it is #41. The next one up is gemma-3-27b-it at 39.7.
gemma-3-12b-it, benchmark by benchmark
gemma-3-12b-it lands below the middle on OTIS Mock AIME: 16.7%, #132 of 189, with 69% of the field scoring higher. Claude Fable 5.1 (max) leads at 100.0%; the closest result above is qwen-plus-2025-01-25 at 17.8%. It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so gemma-3-12b-it is 41.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
gemma-3-12b-it lands below the middle on CritPt: 0.0%, #91 of 115, with 78% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is Gemma 4 26B A4B at 0.0%. It covers unpublished research-level physics problems. The board median is 2.9%, so gemma-3-12b-it is 2.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
GPQA Diamond is a weak spot. gemma-3-12b-it scores 39.5%, #224 of 275, behind 81% of the models measured. The leader, GPT-6 Astra (max), reaches 95.8%. It covers expert-level science questions. The board median is 71.7%, so gemma-3-12b-it is 32.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Chess Puzzles is a weak spot. gemma-3-12b-it scores 0.0%, #101 of 122, behind 82% of the models measured. The leader, GPT-6 Astra (max), reaches 72.0%. It covers chess tactics puzzles. The board median is 12.5%, so gemma-3-12b-it is 12.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
DTBench is a weak spot. gemma-3-12b-it scores 48.8%, #134 of 146, behind 91% of the models measured. The leader, Claude Opus 5.5 (max), reaches 98.9%. It covers dynamic tool-use benchmark. The board median is 68.8%, so gemma-3-12b-it is 20.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
SciCode is a weak spot. gemma-3-12b-it scores 17.4%, #103 of 106, behind 96% of the models measured. The leader, Claude Opus 5.5 (max), reaches 66.9%. It covers scientific research code. The board median is 44.3%, so gemma-3-12b-it is 27.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
LMCA is a weak spot. gemma-3-12b-it scores 4.5%, #107 of 108, behind 98% of the models measured. The leader, Claude Opus 5.5 (max), reaches 68.2%. It covers language-model capability assessment. The board median is 29.2%, so gemma-3-12b-it is 24.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: weak, ahead of 23% of the field on average across 3 benchmarks. Its best board here is OTIS Mock AIME (#132 of 189) and its weakest is GPQA Diamond (#224 of 275). General knowledge and chat: weak, ahead of 9% of the field on average across 2 benchmarks. Its best board here is Chess Puzzles (#101 of 122) and its weakest is LMCA (#107 of 108).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so gemma-3-12b-it costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on gemma-3-27b-it ($0.080) and GPT-4.1 nano ($0.100).
Release history
The Google DeepMind model released before it was Gemini 2.0 Pro Exp (Feb 2025) on 2025-02-05, which scores 46.7. The next one was gemma-3-1b-it on 2025-03-12, at 42.4.
Reading these numbers
The 7 scores on this page come from 1 source: Epoch AI Benchmarking Hub (7). No official leaderboard supplied any of them, so treat them as third-party measurements. With 7 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #736 claude-3-haiku-20240307 (39.9)
- #737 gemma-3-27b-it (39.7)
- #738 Meta-Llama-3-8B-Instruct (39.4)
- #740 GPT-4.1 nano (38.7)
- #741 GPT-3.5 Turbo (Jan 2024) (38.5)
- #742 gpt-4o-mini-2024-07-18 (38.3)
More from Google DeepMind
- Gemini 3.8 Flash (high) (#26, 54.0)
- Gemini 3.7 Flash (high) (#27, 53.9)
- Gemini 3 Pro Preview (#31, 53.7)
- Gemini 3 Deep Think (#60, 52.6)
- AI co-mathematician (#69, 52.3)
- Gemini 3 Flash (#120, 51.5)
Compare it on the live leaderboard → · How the aggregate score works