diffusiongemma-26B-A4B-it: benchmark scores, price and ranking
diffusiongemma-26B-A4B-it from google ranks #202 of 827 LLMs on the LLMs Tiger aggregate benchmark score (50.8), based on 2 benchmark results captured up to 2026-11-18.
- Organisation
- Type
- Open-weight / local model
- Open weights
- Yes
- Aggregate rank
- #202 of 827 (score 50.8)
Across the 2 benchmarks where at least ten models are scored, diffusiongemma-26B-A4B-it is in the top 10% on 0 and in the top half on 1.
diffusiongemma-26B-A4B-it benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| MMLU-Pro | 77.6% | #48 of 158 | HuggingFace MMLU-Pro leaderboard | 2026-11-18 |
| GPQA Diamond | 73.2% | #157 of 302 | HuggingFace GPQA Diamond leaderboard | 2026-11-18 |
About diffusiongemma-26B-A4B-it
diffusiongemma-26B-A4B-it is an open-weight model that can run on your own hardware from google. No API price is listed for it yet. We hold 2 benchmark results for it across 1 area, and none is in the top tenth of its board.
Where diffusiongemma-26B-A4B-it stands
On our aggregate score, diffusiongemma-26B-A4B-it is #202 of 827 ranked models, in the top quarter. Its score is 50.8; the scale is centred on 50, so it sits 0.8 points above the typical model. That beats 76% of the ranked field.
diffusiongemma-26B-A4B-it, benchmark by benchmark
MMLU-Pro puts diffusiongemma-26B-A4B-it at 77.6%, #48 of 158. That beats 70% of the models tested, but 12.2 points short of Intern-S2-Preview-397B (89.8%). The results either side are LongCat-Flash-Lite (78.3%) and LLaDA2.1-flash (76.6%). It covers 57-subject knowledge + reasoning. The board median is 58.1%, so diffusiongemma-26B-A4B-it is 19.5 points above it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-11-18.
diffusiongemma-26B-A4B-it lands below the middle on GPQA Diamond: 73.2%, #157 of 302, with 52% of the field scoring higher. GPT-6.2 Astra (max) leads at 96.8%; the closest result above is Gemma 4 26B A4B at 73.2%. It covers expert-level science questions. The board median is 75.1%, so diffusiongemma-26B-A4B-it is 1.9 points below it. Source: HuggingFace GPQA Diamond leaderboard, captured 2026-11-18.
Strengths and weaknesses
Reasoning, maths and science: solid, ahead of 59% of the field on average across 2 benchmarks. Its best board here is MMLU-Pro (#48 of 158) and its weakest is GPQA Diamond (#157 of 302).
Reading these numbers
The 2 scores on this page come from 2 sources: HuggingFace GPQA Diamond leaderboard (1), HuggingFace MMLU-Pro leaderboard (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 2 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #199 gpt-4-0314 (50.8)
- #200 claude-sonnet-4-20250514_59K (50.8)
- #201 claude-3-5-haiku-20241022 (50.8)
- #203 Gemini 2.5 Pro Preview (Jun 2025) (50.8)
- #204 Claude Code + GLM-5.3 (50.7)
- #205 GPT-5.6 Luna 2026-07-30 (Max) (50.7)
Compare it on the live leaderboard → · How the aggregate score works