Model comparison
Granite 3.0 8b Instruct vs Qwen2.5 7B Instruct
Granite 3.0 8b Instruct is the stronger model overall, scoring 31.1 to 28.6 on the Noometry Index.
Last verified . 2 shared benchmarks.
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. Granite 3.0 8b Instruct scores higher in 3 categories and Qwen2.5 7B Instruct in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Granite 3.0 8b Instruct leads 30.3 to 11.8.
- The biggest single-benchmark swing is BigCodeBench Complete: 35.4% for Granite 3.0 8b Instruct and 46.1% for Qwen2.5 7B Instruct.
Side by side
| Granite 3.0 8b Instruct | Qwen2.5 7B Instruct | |
|---|---|---|
| Provider | IBM | Alibaba (Qwen) |
| Noometry Index | 31.1 | 28.6 |
| Released | — | 2024-09 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $0.17 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 14 | 15 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5 7B Instruct leads
Granite 3.0 8b Instruct: 29.1 (#297), Qwen2.5 7B Instruct: 35.9 (#204)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| BigCodeBench Instruct | 29.3% | 37.6% |
| BigCodeBench Complete | 35.4% | 46.1% |
| LMArena Coding | 1113 | — |
Agentic & Tool Use Not comparable
Granite 3.0 8b Instruct: —, Qwen2.5 7B Instruct: 23.5 (#117)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| BALROG | — | 7.8% |
Reasoning Granite 3.0 8b Instruct leads
Granite 3.0 8b Instruct: 20.0 (#227), Qwen2.5 7B Instruct: 14.3 (#317)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| Chess Puzzles | — | 0% |
| LMArena Hard Prompts | 1092 | — |
| DTBench | — | 47.7% |
| LMCA | — | 6.4% |
| Epoch Capabilities Index | — | 118.44 |
Math Granite 3.0 8b Instruct leads
Granite 3.0 8b Instruct: 30.3 (#211), Qwen2.5 7B Instruct: 11.8 (#300)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 2.5% |
| Omni-MATH | — | 29.4% |
| LMArena Math | 1143 | — |
Knowledge Granite 3.0 8b Instruct leads
Granite 3.0 8b Instruct: 29.1 (#233), Qwen2.5 7B Instruct: 17.0 (#281)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| GPQA Diamond | — | 35.5% |
| MMLU-Pro | — | 53.9% |
| GPQA (HELM) | — | 34.1% |
| LMArena Expert | 1087 | — |
| MMLU | — | 72.9% |
Multilingual Not comparable
Granite 3.0 8b Instruct: 27.4 (#271), Qwen2.5 7B Instruct: —
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Non-English | 1037 | — |
| LMArena Chinese | 1063 | — |
| LMArena Russian | 1060 | — |
Instruction Following Qwen2.5 7B Instruct leads
Granite 3.0 8b Instruct: 56.2 (#271), Qwen2.5 7B Instruct: 63.4 (#226)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| IFEval | — | 74.1% |
| LMArena Instruction Following | 1088 | — |
Long Context Not comparable
Granite 3.0 8b Instruct: 34.0 (#247), Qwen2.5 7B Instruct: —
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Longer Query | 1122 | — |
Writing & Preference Qwen2.5 7B Instruct leads
Granite 3.0 8b Instruct: 31.2 (#280), Qwen2.5 7B Instruct: 48.8 (#194)
| Benchmark | Granite 3.0 8b Instruct | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Text | 1096 | — |
| LMArena Creative Writing | 1071 | — |
| WildBench | — | 73.1% |
| LMArena Multi-Turn | 1063 | — |
Frequently asked questions
Is Granite 3.0 8b Instruct better than Qwen2.5 7B Instruct?
Granite 3.0 8b Instruct is the stronger model overall, scoring 31.1 to 28.6 on the Noometry Index.
Is Granite 3.0 8b Instruct or Qwen2.5 7B Instruct better for coding?
Qwen2.5 7B Instruct scores higher on coding benchmarks: 35.9 versus 29.1 in the Noometry coding category.
How many benchmarks do Granite 3.0 8b Instruct and Qwen2.5 7B Instruct share?
2 benchmarks have published results for both models. Granite 3.0 8b Instruct has 14 scored results on Noometry and Qwen2.5 7B Instruct has 15.