Model comparison
Qwen1.5-72B vs Qwen3-30B-A3B
Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 30.8 on the Noometry Index.
Last verified . 18 shared benchmarks.
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Qwen1.5-72B scores higher in 1 category and Qwen3-30B-A3B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3-30B-A3B leads 41.8 to 11.5.
- The biggest single-benchmark swing is GPQA Diamond: 28.8% for Qwen1.5-72B and 70.1% for Qwen3-30B-A3B.
Side by side
| Qwen1.5-72B | Qwen3-30B-A3B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 30.8 | 38.9 |
| Released | 2024-02-04 | 2025-04-28 |
| Weights | Open | Open |
| Context window | — | 41K |
| Max output | — | 16K |
| Input $ / M tokens | — | $0.12 |
| Output $ / M tokens | — | $0.50 |
| Results tracked | 22 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3-30B-A3B leads
Qwen1.5-72B: 31.9 (#277), Qwen3-30B-A3B: 37.5 (#194)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Coding | 1165 | 1416 |
| SciCode | — | 33.3% |
| WeirdML | — | 29.8% |
| BigCodeBench Instruct | 33.2% | — |
| BigCodeBench Complete | 40.3% | — |
| HumanEval+ | 59.1% | — |
| MBPP+ | 61.6% | — |
Agentic & Tool Use Not comparable
Qwen1.5-72B: —, Qwen3-30B-A3B: 29.8 (#82)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 41.4% |
Reasoning Too close to call
Qwen1.5-72B: 22.2 (#203), Qwen3-30B-A3B: 22.2 (#204)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Hard Prompts | 1148 | 1398 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 8% |
| DTBench | — | 69.3% |
| LMCA | — | 22.4% |
| Epoch Capabilities Index | — | 139.63 |
Math Qwen3-30B-A3B leads
Qwen1.5-72B: 33.2 (#205), Qwen3-30B-A3B: 37.4 (#157)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Math | 1164 | 1394 |
| MathArena Final-Answer Competitions | — | 47.8% |
| OTIS Mock AIME 2024-2025 | — | 70.3% |
Knowledge Qwen3-30B-A3B leads
Qwen1.5-72B: 11.5 (#300), Qwen3-30B-A3B: 41.8 (#105)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| GPQA Diamond | 28.8% | 70.1% |
| LMArena Expert | 1136 | 1396 |
| Confabulations | — | 12.3% |
Multilingual Qwen3-30B-A3B leads
Qwen1.5-72B: 33.2 (#253), Qwen3-30B-A3B: 49.5 (#132)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1135 | 1372 |
| LMArena Chinese | 1186 | 1433 |
| LMArena French | 1159 | 1418 |
| LMArena German | 1084 | 1380 |
| LMArena Japanese | 1061 | 1337 |
| LMArena Korean | 1050 | 1331 |
| LMArena Russian | 1104 | 1370 |
| LMArena Spanish | 1110 | 1404 |
Instruction Following Qwen3-30B-A3B leads
Qwen1.5-72B: 59.3 (#256), Qwen3-30B-A3B: 72.0 (#142)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1141 | 1363 |
Long Context Qwen1.5-72B leads
Qwen1.5-72B: 35.1 (#243), Qwen3-30B-A3B: 31.0 (#283)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Longer Query | 1157 | 1379 |
| Fiction.LiveBench | — | 40.6% |
Writing & Preference Qwen3-30B-A3B leads
Qwen1.5-72B: 37.3 (#258), Qwen3-30B-A3B: 55.6 (#143)
| Benchmark | Qwen1.5-72B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1166 | 1384 |
| LMArena Creative Writing | 1137 | 1317 |
| LMArena Multi-Turn | 1160 | 1378 |
| Short-Story Creative Writing | — | 75.3% |
Frequently asked questions
Is Qwen1.5-72B better than Qwen3-30B-A3B?
Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 30.8 on the Noometry Index.
Is Qwen1.5-72B or Qwen3-30B-A3B better for coding?
Qwen3-30B-A3B scores higher on coding benchmarks: 37.5 versus 31.9 in the Noometry coding category.
How many benchmarks do Qwen1.5-72B and Qwen3-30B-A3B share?
18 benchmarks have published results for both models. Qwen1.5-72B has 22 scored results on Noometry and Qwen3-30B-A3B has 32.