Model comparison
Granite 3.0 8b Instruct vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 31.6 on the Noometry Index.
Last verified . 12 shared benchmarks.
Alibaba (Qwen)Summary
- They share 12 benchmarks with published results for both. Granite 3.0 8b Instruct scores higher in 0 categories and Qwen3.8 27B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.8 27B leads 65.8 to 31.1.
Side by side
| Granite 3.0 8b Instruct | Qwen3.8 27B | |
|---|---|---|
| Provider | IBM | Alibaba (Qwen) |
| Noometry Index | 31.6 | 46.0 |
| Released | — | 2026-08-14 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 33K |
| Input $ / M tokens | — | $0.99 |
| Output $ / M tokens | — | $1.49 |
| Results tracked | 14 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Granite 3.0 8b Instruct: 29.7 (#301), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1112 | 1482 |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
| BigCodeBench Instruct | 29.3% | — |
| BigCodeBench Complete | 35.4% | — |
Agentic & Tool Use Not comparable
Granite 3.0 8b Instruct: —, Qwen3.8 27B: 32.9 (#57)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
Reasoning Qwen3.8 27B leads
Granite 3.0 8b Instruct: 20.9 (#230), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1092 | 1460 |
| ARC-AGI-2 | — | 42.4% |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| DTBench | — | 88% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| Epoch Capabilities Index | — | 149.38 |
Math Qwen3.8 27B leads
Granite 3.0 8b Instruct: 32.8 (#210), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1143 | 1456 |
| ProofBench | — | 16% |
Knowledge Qwen3.8 27B leads
Granite 3.0 8b Instruct: 29.6 (#236), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1087 | 1482 |
Multimodal Not comparable
Granite 3.0 8b Instruct: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
Granite 3.0 8b Instruct: 27.2 (#276), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1037 | 1430 |
| LMArena Chinese | 1063 | 1504 |
| LMArena Russian | 1060 | 1415 |
| LMArena French | — | 1465 |
| LMArena German | — | 1438 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Qwen3.8 27B leads
Granite 3.0 8b Instruct: 56.0 (#276), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1088 | 1439 |
Long Context Qwen3.8 27B leads
Granite 3.0 8b Instruct: 34.0 (#252), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1122 | 1450 |
Writing & Preference Qwen3.8 27B leads
Granite 3.0 8b Instruct: 31.1 (#285), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Granite 3.0 8b Instruct | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1096 | 1441 |
| LMArena Creative Writing | 1071 | 1384 |
| LMArena Multi-Turn | 1063 | 1441 |
| EQ-Bench Creative Writing | — | 1671 |
Frequently asked questions
Is Granite 3.0 8b Instruct better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 31.6 on the Noometry Index.
Is Granite 3.0 8b Instruct or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 29.7 in the Noometry coding category.
How many benchmarks do Granite 3.0 8b Instruct and Qwen3.8 27B share?
12 benchmarks have published results for both models. Granite 3.0 8b Instruct has 14 scored results on Noometry and Qwen3.8 27B has 31.