Model comparison
Llama 3.1-70B vs Qwen3 32B
Qwen3 32B is the stronger model overall, scoring 39.2 to 29.6 on the Noometry Index. Llama 3.1-70B costs 3.1× less per token, which makes it the better buy when Qwen3 32B's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Llama 3.1-70B scores higher in 1 category and Qwen3 32B in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 32B leads 39.7 to 13.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 3.6% for Llama 3.1-70B and 66.9% for Qwen3 32B.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $0.70 / $2.80 for Qwen3 32B.
- Qwen3 32B accepts more context: 131K tokens versus 128K.
Side by side
| Llama 3.1-70B | Qwen3 32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 29.6 | 39.2 |
| Released | 2024-07-23 | 2025-04 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 4K | 16K |
| Input $ / M tokens | $0.40 | $0.70 |
| Output $ / M tokens | $0.40 | $2.80 |
| Results tracked | 35 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 32B leads
Llama 3.1-70B: 30.3 (#296), Qwen3 32B: 37.7 (#190)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Coding | 1260 | 1358 |
| Aider Polyglot | — | 40% |
| SciCode | — | 35.4% |
| WeirdML | 9% | — |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 54.8% | — |
Agentic & Tool Use Qwen3 32B leads
Llama 3.1-70B: 25.1 (#112), Qwen3 32B: 32.6 (#62)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
Reasoning Llama 3.1-70B leads
Llama 3.1-70B: 21.6 (#220), Qwen3 32B: 20.2 (#241)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1334 |
| DTBench | 60% | 67.5% |
| LMCA | 14.8% | 17.3% |
| Epoch Capabilities Index | 125.92 | 138.51 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
Math Qwen3 32B leads
Llama 3.1-70B: 13.5 (#304), Qwen3 32B: 39.7 (#99)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 3.6% | 66.9% |
| LMArena Math | 1252 | 1399 |
| Omni-MATH | 21% | — |
| MATH Level 5 | 36.7% | — |
Knowledge Qwen3 32B leads
Llama 3.1-70B: 24.2 (#269), Qwen3 32B: 40.0 (#125)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 44.2% | 65.7% |
| LMArena Expert | 1209 | 1362 |
| MMLU-Pro | 65.3% | — |
| Vectara Hallucination Rate | — | 5.9% |
| GPQA (HELM) | 42.6% | — |
| MMLU | 80.1% | — |
Multilingual Qwen3 32B leads
Llama 3.1-70B: 38.8 (#225), Qwen3 32B: 45.6 (#167)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1219 | 1317 |
| LMArena Chinese | 1215 | 1357 |
| LMArena German | 1222 | 1341 |
| LMArena Russian | 1234 | 1311 |
| LMArena French | 1261 | — |
| LMArena Japanese | 1132 | — |
| LMArena Korean | 1140 | — |
| LMArena Spanish | 1253 | — |
Instruction Following Qwen3 32B leads
Llama 3.1-70B: 65.3 (#223), Qwen3 32B: 68.9 (#179)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1231 | 1305 |
| IFEval | 82.1% | — |
Long Context Qwen3 32B leads
Llama 3.1-70B: 37.6 (#214), Qwen3 32B: 43.8 (#87)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1241 | 1327 |
| Fiction.LiveBench | — | 74.2% |
Writing & Preference Qwen3 32B leads
Llama 3.1-70B: 35.4 (#267), Qwen3 32B: 52.9 (#163)
| Benchmark | Llama 3.1-70B | Qwen3 32B |
|---|---|---|
| LMArena Text | 1261 | 1340 |
| LMArena Creative Writing | 1232 | 1297 |
| LMArena Multi-Turn | 1256 | 1331 |
| EQ-Bench Creative Writing | 784 | — |
| WildBench | 75.8% | — |
Frequently asked questions
Is Llama 3.1-70B better than Qwen3 32B?
Qwen3 32B is the stronger model overall, scoring 39.2 to 29.6 on the Noometry Index. Llama 3.1-70B costs 3.1× less per token, which makes it the better buy when Qwen3 32B's lead doesn't matter for your workload.
Which is cheaper, Llama 3.1-70B or Qwen3 32B?
Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.
Is Llama 3.1-70B or Qwen3 32B better for coding?
Qwen3 32B scores higher on coding benchmarks: 37.7 versus 30.3 in the Noometry coding category.
Which has the bigger context window?
Qwen3 32B does, with 131K tokens against 128K.
How many benchmarks do Llama 3.1-70B and Qwen3 32B share?
18 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and Qwen3 32B has 26.