# Llama 4 Maverick vs Qwen3 32B

> Qwen3 32B is the stronger model overall, scoring 39.2 to 30.9 on the Noometry Index. Llama 4 Maverick costs 4.0× less per token, which makes it the better buy when Qwen3 32B's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/llama-4-maverick-vs-qwen3-32b
- Last updated: 2026-10-12
- Shared benchmarks: 25

## Summary

- They share 25 benchmarks with published results for both. Llama 4 Maverick scores higher in 1 category and Qwen3 32B in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3 32B leads 52.9 to 38.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 20.6% for Llama 4 Maverick and 66.9% for Qwen3 32B.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $0.70 / $2.80 for Qwen3 32B.
- Qwen3 32B accepts more context: 131K tokens versus 128K.

## Snapshot

| | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 30.9 | 39.2 |
| Rank | 282 | 172 |
| Context | 128K | 131K |
| Input $/M | $0.19 | $0.70 |
| Output $/M | $0.65 | $2.80 |
| Weights | Open | Open |

## Coding

- Llama 4 Maverick: 26.6 (#324)
- Qwen3 32B: 37.7 (#190)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| Aider Polyglot | 15.6% | 40% |
| SciCode | 33.1% | 35.4% |
| LMArena Coding | 1302 | 1358 |
| SWE-bench Verified (bash only) | 21% | — |
| WeirdML | 24.5% | — |
| BigCodeBench Instruct | 49.7% | — |
| BigCodeBench Complete | 61.4% | — |
| ALE-Bench | 172.97 | — |

## Agentic & Tool Use

- Llama 4 Maverick: 28.2 (#91)
- Qwen3 32B: 32.6 (#62)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.3% | 48.7% |

## Reasoning

- Llama 4 Maverick: 10.1 (#342)
- Qwen3 32B: 20.2 (#241)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | 55.9% | 54.9% |
| CritPt | 0% | 0.3% |
| LMArena Hard Prompts | 1281 | 1334 |
| DTBench | 61.9% | 67.5% |
| LMCA | 15.9% | 17.3% |
| Epoch Capabilities Index | 132.2 | 138.51 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 27.7% | — |
| NYT Connections (extended) | 8% | — |
| ARC-AGI-1 | 4.4% | — |
| Chess Puzzles | — | 5% |
| EnigmaEval | 0.6% | — |
| ForecastBench | 57.5 | — |

## Math

- Llama 4 Maverick: 26.0 (#262)
- Qwen3 32B: 39.7 (#99)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 20.6% | 66.9% |
| LMArena Math | 1299 | 1399 |
| Omni-MATH | 42.2% | — |
| MATH Level 5 | 73% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |

## Knowledge

- Llama 4 Maverick: 33.4 (#204)
- Qwen3 32B: 40.0 (#125)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 67% | 65.7% |
| Vectara Hallucination Rate | 8.2% | 5.9% |
| LMArena Expert | 1259 | 1362 |
| Humanity's Last Exam | 5.7% | — |
| MMLU-Pro | 81% | — |
| Confabulations | 22.6% | — |
| GPQA (HELM) | 65% | — |

## Multimodal

- Llama 4 Maverick: 31.6 (#105)
- Qwen3 32B: —

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| LMArena Vision | 1142 | — |
| GeoBench | 52% | — |
| SpatialViz-Bench | 31.8% | — |

## Multilingual

- Llama 4 Maverick: 42.2 (#195)
- Qwen3 32B: 45.6 (#167)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1269 | 1317 |
| LMArena Chinese | 1277 | 1357 |
| LMArena German | 1291 | 1341 |
| LMArena Russian | 1286 | 1311 |
| LMArena French | 1259 | — |
| LMArena Japanese | 1207 | — |
| LMArena Korean | 1203 | — |
| LMArena Spanish | 1293 | — |

## Instruction Following

- Llama 4 Maverick: 71.7 (#146)
- Qwen3 32B: 68.9 (#179)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1267 | 1305 |
| IFEval | 90.8% | — |

## Long Context

- Llama 4 Maverick: 31.4 (#279)
- Qwen3 32B: 43.8 (#87)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | 46.2% | 74.2% |
| LMArena Longer Query | 1280 | 1327 |

## Writing & Preference

- Llama 4 Maverick: 38.8 (#252)
- Qwen3 32B: 52.9 (#163)

| Benchmark | Llama 4 Maverick | Qwen3 32B |
|---|---|---|
| LMArena Text | 1287 | 1340 |
| LMArena Creative Writing | 1267 | 1297 |
| LMArena Multi-Turn | 1289 | 1331 |
| Short-Story Creative Writing | 62% | — |
| EQ-Bench Creative Writing | 860 | — |
| WildBench | 80% | — |

## FAQ

### Is Llama 4 Maverick better than Qwen3 32B?

Qwen3 32B is the stronger model overall, scoring 39.2 to 30.9 on the Noometry Index. Llama 4 Maverick costs 4.0× 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 Maverick or Qwen3 32B?

Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.

### Is Llama 4 Maverick or Qwen3 32B better for coding?

Qwen3 32B scores higher on coding benchmarks: 37.7 versus 26.6 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 Maverick and Qwen3 32B share?

25 benchmarks have published results for both models. Llama 4 Maverick has 54 scored results on Noometry and Qwen3 32B has 26.