Model comparison
Granite 4.2 3b vs Qwen2.5 72B Instruct
Granite 4.2 3b is the stronger model overall, scoring 39.4 to 31.9 on the Noometry Index.
Last verified . 11 shared benchmarks.
Alibaba (Qwen)Summary
- They share 11 benchmarks with published results for both. Granite 4.2 3b scores higher in 7 categories and Qwen2.5 72B Instruct in 0 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Granite 4.2 3b leads 36.3 to 27.0.
Side by side
| Granite 4.2 3b | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | IBM | Alibaba (Qwen) |
| Noometry Index | 39.4 | 31.9 |
| Released | — | 2024-09 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $1.40 |
| Output $ / M tokens | — | $5.60 |
| Results tracked | 11 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Granite 4.2 3b leads
Granite 4.2 3b: 40.0 (#151), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1361 | 1292 |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
Agentic & Tool Use Not comparable
Granite 4.2 3b: —, Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Granite 4.2 3b leads
Granite 4.2 3b: 26.0 (#138), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1306 | 1271 |
| DTBench | — | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Not comparable
Granite 4.2 3b: —, Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 8.1% |
| Omni-MATH | — | 33% |
| LMArena Math | — | 1283 |
| MATH Level 5 | — | 63.2% |
Knowledge Granite 4.2 3b leads
Granite 4.2 3b: 36.3 (#171), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Expert | 1315 | 1245 |
| GPQA Diamond | — | 49.1% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multilingual Granite 4.2 3b leads
Granite 4.2 3b: 42.1 (#198), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1268 | 1252 |
| LMArena Chinese | 1269 | 1272 |
| LMArena Russian | 1249 | 1264 |
| LMArena French | — | 1280 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Spanish | — | 1256 |
Instruction Following Granite 4.2 3b leads
Granite 4.2 3b: 67.1 (#200), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1273 | 1254 |
| IFEval | — | 80.6% |
Long Context Too close to call
Granite 4.2 3b: 39.2 (#185), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1291 | 1282 |
Writing & Preference Too close to call
Granite 4.2 3b: 47.2 (#212), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | Granite 4.2 3b | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1293 | 1269 |
| LMArena Creative Writing | 1205 | 1221 |
| LMArena Multi-Turn | 1290 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is Granite 4.2 3b better than Qwen2.5 72B Instruct?
Granite 4.2 3b is the stronger model overall, scoring 39.4 to 31.9 on the Noometry Index.
Is Granite 4.2 3b or Qwen2.5 72B Instruct better for coding?
Granite 4.2 3b scores higher on coding benchmarks: 40.0 versus 33.2 in the Noometry coding category.
How many benchmarks do Granite 4.2 3b and Qwen2.5 72B Instruct share?
11 benchmarks have published results for both models. Granite 4.2 3b has 11 scored results on Noometry and Qwen2.5 72B Instruct has 43.