Model comparison
Llama-3.3-70B-Instruct vs Qwen3-30B-A3B
Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 30.6 on the Noometry Index.
Last verified . 28 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 28 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 0 categories and Qwen3-30B-A3B in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3-30B-A3B leads 37.4 to 15.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 5.1% for Llama-3.3-70B-Instruct and 70.3% for Qwen3-30B-A3B.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.12 / $0.50 for Qwen3-30B-A3B.
- Llama-3.3-70B-Instruct accepts more context: 128K tokens versus 41K.
Side by side
| Llama-3.3-70B-Instruct | Qwen3-30B-A3B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 30.6 | 38.9 |
| Released | 2024-12-06 | 2025-04-28 |
| Weights | Open | Open |
| Context window | 128K | 41K |
| Max output | 4K | 16K |
| Input $ / M tokens | $0.10 | $0.12 |
| Output $ / M tokens | $0.32 | $0.50 |
| Results tracked | 43 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 31.0 (#290), Qwen3-30B-A3B: 37.5 (#194)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| SciCode | 26% | 33.3% |
| WeirdML | 14.4% | 29.8% |
| LMArena Coding | 1268 | 1416 |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
Agentic & Tool Use Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 25.8 (#105), Qwen3-30B-A3B: 29.8 (#82)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | 41.4% |
| BALROG | 23% | — |
Reasoning Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 14.1 (#327), Qwen3-30B-A3B: 22.2 (#204)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| CritPt | 0% | 0.3% |
| LMArena Hard Prompts | 1257 | 1398 |
| DTBench | 59.5% | 69.3% |
| LMCA | 17.5% | 22.4% |
| Epoch Capabilities Index | 127.33 | 139.63 |
| SimpleBench | 19.9% | — |
| Kagi LLM Benchmark | — | 54.9% |
| Chess Puzzles | — | 8% |
| LiveBench Reasoning | 50.8% | — |
| LiveBench Data Analysis | 49.5% | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |
Math Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 15.3 (#298), Qwen3-30B-A3B: 37.4 (#157)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 5.1% | 70.3% |
| LMArena Math | 1267 | 1394 |
| MathArena Final-Answer Competitions | — | 47.8% |
| LiveBench Math | 42.2% | — |
| MATH Level 5 | 41.6% | — |
Knowledge Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 30.6 (#226), Qwen3-30B-A3B: 41.8 (#105)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| GPQA Diamond | 47.4% | 70.1% |
| Confabulations | 22.8% | 12.3% |
| LMArena Expert | 1225 | 1396 |
| Vectara Hallucination Rate | 4.1% | — |
| MMLU | 86.3% | — |
Multilingual Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 39.9 (#220), Qwen3-30B-A3B: 49.5 (#132)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1236 | 1372 |
| LMArena Chinese | 1217 | 1433 |
| LMArena French | 1281 | 1418 |
| LMArena German | 1251 | 1380 |
| LMArena Japanese | 1150 | 1337 |
| LMArena Korean | 1143 | 1331 |
| LMArena Russian | 1252 | 1370 |
| LMArena Spanish | 1270 | 1404 |
Instruction Following Too close to call
Llama-3.3-70B-Instruct: 71.1 (#157), Qwen3-30B-A3B: 72.0 (#142)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1242 | 1363 |
| LiveBench Instruction Following | 82.7% | — |
Long Context Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 26.4 (#295), Qwen3-30B-A3B: 31.0 (#283)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| Fiction.LiveBench | 33.3% | 40.6% |
| LMArena Longer Query | 1256 | 1379 |
Writing & Preference Qwen3-30B-A3B leads
Llama-3.3-70B-Instruct: 47.6 (#207), Qwen3-30B-A3B: 55.6 (#143)
| Benchmark | Llama-3.3-70B-Instruct | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1274 | 1384 |
| LMArena Creative Writing | 1250 | 1317 |
| LMArena Multi-Turn | 1280 | 1378 |
| Short-Story Creative Writing | — | 75.3% |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than Qwen3-30B-A3B?
Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 30.6 on the Noometry Index.
Which is cheaper, Llama-3.3-70B-Instruct or Qwen3-30B-A3B?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Qwen3-30B-A3B lists at $0.12 and $0.50.
Is Llama-3.3-70B-Instruct or Qwen3-30B-A3B better for coding?
Qwen3-30B-A3B scores higher on coding benchmarks: 37.5 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
Llama-3.3-70B-Instruct does, with 128K tokens against 41K.
How many benchmarks do Llama-3.3-70B-Instruct and Qwen3-30B-A3B share?
28 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Qwen3-30B-A3B has 32.