granite-4.1-30b: benchmark scores, price and ranking
granite-4.1-30b from ibm-granite ranks #577 of 743 LLMs on the LLMs Tiger aggregate benchmark score (45.5), based on 5 benchmark results captured up to 2026-10-13.
- Organisation
- ibm-granite
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2026-04-29 (source: epoch)
- Aggregate rank
- #577 of 743 (score 45.5)
Relative to other models, granite-4.1-30b is strongest on GSM8K (#3 of 17) and weakest on SciCode (#99 of 106).
Across the 5 benchmarks where at least ten models are scored, granite-4.1-30b is in the top 10% on 0 and in the top half on 2.
granite-4.1-30b benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| GSM8K | 94.2% | #3 of 17 | HuggingFace GSM8K leaderboard | 2026-10-13 |
| MMLU-Pro | 64.1% | #54 of 136 | HuggingFace MMLU-Pro leaderboard | 2026-10-13 |
| GPQA Diamond | 45.8% | #208 of 275 | HuggingFace GPQA Diamond leaderboard | 2026-10-13 |
| CritPt | 0.0% | #98 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 25.8% | #99 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
About granite-4.1-30b
granite-4.1-30b is an open-weight model that can run on your own hardware from ibm-granite. It was released on 2026-04-29, according to Epoch AI. No API price is listed for it yet. We hold 5 benchmark results for it across 3 areas, and none is in the top tenth of its board.
Where granite-4.1-30b stands
On our aggregate score, granite-4.1-30b is #577 of 743 ranked models, in the lower half. Its score is 45.5; the scale is centred on 50, so it sits 4.5 points below the typical model. That beats 22% of the ranked field. Among the 9 ranked ibm-granite models it is #2. The next one up is granite-4.1-8b at 47.1. The next one down is granite-3.1-8b-base at 44.9.
granite-4.1-30b, benchmark by benchmark
GSM8K puts granite-4.1-30b at 94.2%, #3 of 17. That beats 82% of the models tested, but 5.4 points short of mimo-v2.5-pro (99.6%). The results either side are Llama 3.1-405B (96.8%) and Llama-3.2-90B-Vision-Instruct (93.1%). It covers grade-school math word problems. The board median is 86.9%, so granite-4.1-30b is 7.3 points above it. Source: HuggingFace GSM8K leaderboard, captured 2026-10-13.
On MMLU-Pro, granite-4.1-30b reaches 64.1%. It ranks #54 of 136, ahead of 60% of the field, and 24.0 points behind Darwin-180B-RSI at 88.1%. The results either side are DeepSeek-V3 (64.4%) and Qwen2.5-14B (63.7%). It covers 57-subject knowledge + reasoning. The board median is 55.3%, so granite-4.1-30b is 8.8 points above it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-10-13.
granite-4.1-30b lands below the middle on GPQA Diamond: 45.8%, #208 of 275, with 75% of the field scoring higher. GPT-6 Astra (max) leads at 95.8%; the closest result above is gemini-1.5-pro-001 at 45.9%. It covers expert-level science questions. The board median is 71.7%, so granite-4.1-30b is 26.0 points below it. Source: HuggingFace GPQA Diamond leaderboard, captured 2026-10-13.
CritPt is a weak spot. granite-4.1-30b scores 0.0%, #98 of 115, behind 84% of the models measured. The leader, GPT-5.6 Sol (max), reaches 32.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so granite-4.1-30b is 2.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
SciCode is a weak spot. granite-4.1-30b scores 25.8%, #99 of 106, behind 92% 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 granite-4.1-30b is 18.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: uneven, ahead of 33% of the field on average across 3 benchmarks. Its best board here is MMLU-Pro (#54 of 136) and its weakest is CritPt (#98 of 115).
Reading these numbers
The 5 scores on this page come from 4 sources: Epoch AI Benchmarking Hub (2), HuggingFace GPQA Diamond leaderboard (1), HuggingFace GSM8K leaderboard (1), HuggingFace MMLU-Pro leaderboard (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 5 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #574 Yi-1.5-6B-Chat (45.6)
- #575 qwen3-coder-next (45.5)
- #576 Qwen1.5-14B-Chat (45.5)
- #578 Ministral-8B-Instruct-2410 (45.5)
- #579 c4ai-command-r-v01 (45.5)
- #580 qwen1.5-72b-chat (45.5)
More from ibm-granite
- granite-4.1-8b (#461, 47.1)
- granite-3.1-8b-base (#619, 44.9)
- granite-3.0-8b-base (#629, 44.6)
- granite-4.1-3b (#639, 44.4)
- granite-3.1-2b-base (#669, 43.6)
- granite-3.0-2b-base (#681, 43.3)
Compare it on the live leaderboard → · How the aggregate score works