Model comparison
Granite 4.0 Micro vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 28.8 on the Noometry Index. Granite 4.0 Micro costs 60× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Last verified . 7 shared benchmarks.
Alibaba (Qwen)Summary
- They share 7 benchmarks with published results for both. Granite 4.0 Micro scores higher in 1 category and Qwen2.5 72B Instruct in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen2.5 72B Instruct leads 26.9 to 9.9.
- The biggest single-benchmark swing is MMLU-Pro: 39.5% for Granite 4.0 Micro and 63.1% for Qwen2.5 72B Instruct.
- Granite 4.0 Micro is cheaper at $0.017 / $0.11 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct accepts more context: 131K tokens versus 131K.
Side by side
| Granite 4.0 Micro | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | IBM | Alibaba (Qwen) |
| Noometry Index | 28.8 | 31.5 |
| Released | 2025-10-02 | 2024-09 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 118K | 8K |
| Input $ / M tokens | $0.017 | $1.40 |
| Output $ / M tokens | $0.11 | $5.60 |
| Results tracked | 8 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Granite 4.0 Micro: —, Qwen2.5 72B Instruct: 32.8 (#256)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| LMArena Coding | — | 1293 |
| BigCodeBench Complete | — | 55.9% |
Agentic & Tool Use Not comparable
Granite 4.0 Micro: —, Qwen2.5 72B Instruct: 22.0 (#124)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Granite 4.0 Micro: 18.5 (#261), Qwen2.5 72B Instruct: 21.4 (#197)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| Chess Puzzles | 0% | — |
| LMArena Hard Prompts | — | 1271 |
| DTBench | — | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129.01 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Qwen2.5 72B Instruct leads
Granite 4.0 Micro: 11.2 (#301), Qwen2.5 72B Instruct: 18.1 (#281)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.8% | 8.1% |
| Omni-MATH | 20.9% | 33% |
| LMArena Math | — | 1283 |
| MATH Level 5 | — | 63.2% |
Knowledge Qwen2.5 72B Instruct leads
Granite 4.0 Micro: 9.9 (#299), Qwen2.5 72B Instruct: 26.9 (#247)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 28.3% | 49.1% |
| MMLU-Pro | 39.5% | 63.1% |
| GPQA (HELM) | 30.7% | 42.6% |
| Confabulations | — | 19.1% |
| LMArena Expert | — | 1245 |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multilingual Not comparable
Granite 4.0 Micro: —, Qwen2.5 72B Instruct: 41.2 (#208)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | — | 1252 |
| LMArena Chinese | — | 1272 |
| LMArena French | — | 1279 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Russian | — | 1264 |
| LMArena Spanish | — | 1255 |
Instruction Following Granite 4.0 Micro leads
Granite 4.0 Micro: 70.0 (#164), Qwen2.5 72B Instruct: 65.7 (#216)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| IFEval | 84.9% | 80.6% |
| LMArena Instruction Following | — | 1254 |
Long Context Not comparable
Granite 4.0 Micro: —, Qwen2.5 72B Instruct: 38.9 (#183)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | — | 1282 |
Writing & Preference Too close to call
Granite 4.0 Micro: 46.7 (#212), Qwen2.5 72B Instruct: 46.9 (#210)
| Benchmark | Granite 4.0 Micro | Qwen2.5 72B Instruct |
|---|---|---|
| WildBench | 67% | 80.2% |
| LMArena Text | — | 1269 |
| LMArena Creative Writing | — | 1221 |
| LMArena Multi-Turn | — | 1272 |
Frequently asked questions
Is Granite 4.0 Micro better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 28.8 on the Noometry Index. Granite 4.0 Micro costs 60× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Which is cheaper, Granite 4.0 Micro or Qwen2.5 72B Instruct?
Granite 4.0 Micro is cheaper. It lists at $0.017 per million input tokens and $0.11 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Which has the bigger context window?
Qwen2.5 72B Instruct does, with 131K tokens against 131K.
How many benchmarks do Granite 4.0 Micro and Qwen2.5 72B Instruct share?
7 benchmarks have published results for both models. Granite 4.0 Micro has 8 scored results on Noometry and Qwen2.5 72B Instruct has 43.