Model comparison
Llama 3.2 3B vs Qwen3.5 122B-A10B
Qwen3.5 122B-A10B is the stronger model overall, scoring 42.1 to 28.9 on the Noometry Index. Llama 3.2 3B costs 9.2× less per token, which makes it the better buy when Qwen3.5 122B-A10B's lead doesn't matter for your workload.
Last verified . 13 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. Llama 3.2 3B scores higher in 0 categories and Qwen3.5 122B-A10B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.5 122B-A10B leads 60.0 to 24.7.
- Llama 3.2 3B is cheaper at $0.05 / $0.33 per million input/output tokens, against $0.40 / $3.20 for Qwen3.5 122B-A10B.
- Qwen3.5 122B-A10B accepts more context: 262K tokens versus 131K.
Side by side
| Llama 3.2 3B | Qwen3.5 122B-A10B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.9 | 42.1 |
| Released | 2024-09-24 | 2026-02-23 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 118K | 66K |
| Input $ / M tokens | $0.05 | $0.40 |
| Output $ / M tokens | $0.33 | $3.20 |
| Results tracked | 18 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 122B-A10B leads
Llama 3.2 3B: 27.6 (#319), Qwen3.5 122B-A10B: 39.1 (#162)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Coding | 1098 | 1436 |
| LMArena WebDev | — | 1360 |
| SciCode | — | 35.6% |
| BigCodeBench Instruct | 23.4% | — |
| BigCodeBench Complete | 28.3% | — |
Agentic & Tool Use Not comparable
Llama 3.2 3B: 20.1 (#143), Qwen3.5 122B-A10B: —
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 21.9% | — |
| BALROG | 10.1% | — |
Reasoning Qwen3.5 122B-A10B leads
Llama 3.2 3B: 21.0 (#228), Qwen3.5 122B-A10B: 27.2 (#123)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Hard Prompts | 1095 | 1421 |
| NYT Connections (extended) | — | 51.7% |
| CritPt | — | 0.9% |
| Thematic Generalization | — | 51.2% |
| Mystery Game Puzzles | — | 17% |
| DTBench | — | 84.3% |
| LMCA | — | 32.2% |
Math Qwen3.5 122B-A10B leads
Llama 3.2 3B: 32.4 (#214), Qwen3.5 122B-A10B: 39.1 (#112)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Math | 1126 | 1432 |
Knowledge Qwen3.5 122B-A10B leads
Llama 3.2 3B: 29.7 (#235), Qwen3.5 122B-A10B: 38.8 (#142)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Expert | 1090 | 1432 |
| Vectara Hallucination Rate | — | 11.2% |
Multimodal Not comparable
Llama 3.2 3B: —, Qwen3.5 122B-A10B: 39.6 (#57)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Vision | — | 1245 |
Multilingual Qwen3.5 122B-A10B leads
Llama 3.2 3B: 26.2 (#281), Qwen3.5 122B-A10B: 51.6 (#107)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Non-English | 1019 | 1400 |
| LMArena Chinese | 1017 | 1462 |
| LMArena German | 1056 | 1426 |
| LMArena Russian | 949 | 1400 |
| LMArena French | — | 1442 |
| LMArena Japanese | — | 1367 |
| LMArena Korean | — | 1352 |
| LMArena Spanish | — | 1424 |
Instruction Following Qwen3.5 122B-A10B leads
Llama 3.2 3B: 56.0 (#275), Qwen3.5 122B-A10B: 73.8 (#115)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Instruction Following | 1089 | 1399 |
Long Context Qwen3.5 122B-A10B leads
Llama 3.2 3B: 33.4 (#261), Qwen3.5 122B-A10B: 43.0 (#109)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Longer Query | 1100 | 1410 |
Writing & Preference Qwen3.5 122B-A10B leads
Llama 3.2 3B: 24.7 (#307), Qwen3.5 122B-A10B: 60.0 (#105)
| Benchmark | Llama 3.2 3B | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Text | 1110 | 1417 |
| LMArena Creative Writing | 1094 | 1368 |
| LMArena Multi-Turn | 1105 | 1416 |
| EQ-Bench Creative Writing | 595 | — |
Frequently asked questions
Is Llama 3.2 3B better than Qwen3.5 122B-A10B?
Qwen3.5 122B-A10B is the stronger model overall, scoring 42.1 to 28.9 on the Noometry Index. Llama 3.2 3B costs 9.2× less per token, which makes it the better buy when Qwen3.5 122B-A10B's lead doesn't matter for your workload.
Which is cheaper, Llama 3.2 3B or Qwen3.5 122B-A10B?
Llama 3.2 3B is cheaper. It lists at $0.05 per million input tokens and $0.33 per million output tokens; Qwen3.5 122B-A10B lists at $0.40 and $3.20.
Is Llama 3.2 3B or Qwen3.5 122B-A10B better for coding?
Qwen3.5 122B-A10B scores higher on coding benchmarks: 39.1 versus 27.6 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5 122B-A10B does, with 262K tokens against 131K.
How many benchmarks do Llama 3.2 3B and Qwen3.5 122B-A10B share?
13 benchmarks have published results for both models. Llama 3.2 3B has 18 scored results on Noometry and Qwen3.5 122B-A10B has 27.