Model comparison
Llama 4 Scout vs Qwen3 32B
Qwen3 32B is the stronger model overall, scoring 39.2 to 27.7 on the Noometry Index. Llama 4 Scout costs 8.2× less per token, which makes it the better buy when Qwen3 32B's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. Llama 4 Scout scores higher in 0 categories and Qwen3 32B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 32B leads 39.7 to 19.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.8% for Llama 4 Scout and 66.9% for Qwen3 32B.
- Llama 4 Scout is cheaper at $0.10 / $0.30 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 4 Scout | Qwen3 32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 27.7 | 39.2 |
| Released | 2025-04-05 | 2025-04 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 4K | 16K |
| Input $ / M tokens | $0.10 | $0.70 |
| Output $ / M tokens | $0.30 | $2.80 |
| Results tracked | 43 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 32B leads
Llama 4 Scout: 20.2 (#339), Qwen3 32B: 37.7 (#190)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| SciCode | 17% | 35.4% |
| LMArena Coding | 1286 | 1358 |
| SWE-bench Verified (bash only) | 9.1% | — |
| Aider Polyglot | — | 40% |
| BigCodeBench Complete | 43.1% | — |
Agentic & Tool Use Qwen3 32B leads
Llama 4 Scout: 24.6 (#119), Qwen3 32B: 32.6 (#62)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 28.1% | 48.7% |
Reasoning Qwen3 32B leads
Llama 4 Scout: 9.1 (#345), Qwen3 32B: 20.2 (#241)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | 36.9% | 54.9% |
| CritPt | 0% | 0.3% |
| LMArena Hard Prompts | 1266 | 1334 |
| DTBench | 57.9% | 67.5% |
| LMCA | 12% | 17.3% |
| Epoch Capabilities Index | 129.64 | 138.51 |
| ARC-AGI-2 | 0% | — |
| ARC-AGI-1 | 0.5% | — |
| Chess Puzzles | — | 5% |
| ForecastBench | 57.5 | — |
Math Qwen3 32B leads
Llama 4 Scout: 19.6 (#286), Qwen3 32B: 39.7 (#99)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.8% | 66.9% |
| LMArena Math | 1287 | 1399 |
| Omni-MATH | 37.3% | — |
| MATH Level 5 | 62.3% | — |
| FrontierMath (Feb 2025 set) | 0% | — |
Knowledge Qwen3 32B leads
Llama 4 Scout: 31.9 (#217), Qwen3 32B: 40.0 (#125)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 51.8% | 65.7% |
| Vectara Hallucination Rate | 7.7% | 5.9% |
| LMArena Expert | 1235 | 1362 |
| MMLU-Pro | 74.2% | — |
| GPQA (HELM) | 50.7% | — |
Multimodal Not comparable
Llama 4 Scout: 32.2 (#102), Qwen3 32B: —
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| LMArena Vision | 1118 | — |
| SpatialViz-Bench | 34.2% | — |
Multilingual Qwen3 32B leads
Llama 4 Scout: 41.0 (#212), Qwen3 32B: 45.6 (#167)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1252 | 1317 |
| LMArena Chinese | 1255 | 1357 |
| LMArena German | 1272 | 1341 |
| LMArena Russian | 1263 | 1311 |
| LMArena French | 1282 | — |
| LMArena Japanese | 1206 | — |
| LMArena Korean | 1207 | — |
| LMArena Spanish | 1278 | — |
Instruction Following Qwen3 32B leads
Llama 4 Scout: 65.8 (#217), Qwen3 32B: 68.9 (#179)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1248 | 1305 |
| IFEval | 81.8% | — |
Long Context Qwen3 32B leads
Llama 4 Scout: 27.5 (#294), Qwen3 32B: 43.8 (#87)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | 36% | 74.2% |
| LMArena Longer Query | 1265 | 1327 |
Writing & Preference Qwen3 32B leads
Llama 4 Scout: 37.0 (#261), Qwen3 32B: 52.9 (#163)
| Benchmark | Llama 4 Scout | Qwen3 32B |
|---|---|---|
| LMArena Text | 1279 | 1340 |
| LMArena Creative Writing | 1249 | 1297 |
| LMArena Multi-Turn | 1280 | 1331 |
| EQ-Bench Creative Writing | 783 | — |
| WildBench | 78% | — |
Frequently asked questions
Is Llama 4 Scout better than Qwen3 32B?
Qwen3 32B is the stronger model overall, scoring 39.2 to 27.7 on the Noometry Index. Llama 4 Scout costs 8.2× less per token, which makes it the better buy when Qwen3 32B's lead doesn't matter for your workload.
Which is cheaper, Llama 4 Scout or Qwen3 32B?
Llama 4 Scout is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.
Is Llama 4 Scout or Qwen3 32B better for coding?
Qwen3 32B scores higher on coding benchmarks: 37.7 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
Qwen3 32B does, with 131K tokens against 128K.
How many benchmarks do Llama 4 Scout and Qwen3 32B share?
24 benchmarks have published results for both models. Llama 4 Scout has 43 scored results on Noometry and Qwen3 32B has 26.