Model comparison
Amazon Nova Lite vs Qwen3.5 122B-A10B
Qwen3.5 122B-A10B is the stronger model overall, scoring 41.6 to 31.4 on the Noometry Index. Amazon Nova Lite costs 10× less per token, which makes it the better buy when Qwen3.5 122B-A10B's lead doesn't matter for your workload.
Last verified . 19 shared benchmarks.
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Amazon Nova Lite scores higher in 0 categories and Qwen3.5 122B-A10B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.5 122B-A10B leads 60.2 to 42.3.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 6.1% for Amazon Nova Lite and 11.2% for Qwen3.5 122B-A10B.
- Amazon Nova Lite is cheaper at $0.06 / $0.24 per million input/output tokens, against $0.40 / $3.20 for Qwen3.5 122B-A10B.
- Amazon Nova Lite accepts more context: 300K tokens versus 262K.
- Qwen3.5 122B-A10B has downloadable open weights; the other is API-only.
Side by side
| Amazon Nova Lite | Qwen3.5 122B-A10B | |
|---|---|---|
| Provider | Amazon | Alibaba (Qwen) |
| Noometry Index | 31.4 | 41.6 |
| Released | 2024-12-03 | 2026-02-23 |
| Weights | Proprietary | Open |
| Context window | 300K | 262K |
| Max output | 10K | 66K |
| Input $ / M tokens | $0.06 | $0.40 |
| Output $ / M tokens | $0.24 | $3.20 |
| Results tracked | 34 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 122B-A10B leads
Amazon Nova Lite: 32.0 (#266), Qwen3.5 122B-A10B: 38.8 (#155)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Coding | 1239 | 1436 |
| LMArena WebDev | — | 1360 |
| SciCode | — | 35.6% |
| LiveBench Coding | 27.5% | — |
| ALE-Bench | 236.25 | — |
Reasoning Qwen3.5 122B-A10B leads
Amazon Nova Lite: 18.5 (#257), Qwen3.5 122B-A10B: 26.4 (#121)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1421 |
| NYT Connections (extended) | — | 51.7% |
| CritPt | — | 0.9% |
| Thematic Generalization | — | 51.2% |
| LiveBench Reasoning | 36.7% | — |
| Mystery Game Puzzles | — | 17% |
| DTBench | — | 84.3% |
| LiveBench Data Analysis | 37.2% | — |
| LMCA | — | 32.2% |
| LiveBench | 36.4% | — |
Math Qwen3.5 122B-A10B leads
Amazon Nova Lite: 25.8 (#243), Qwen3.5 122B-A10B: 36.4 (#114)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Math | 1227 | 1432 |
| Omni-MATH | 23.3% | — |
| LiveBench Math | 36.7% | — |
Knowledge Qwen3.5 122B-A10B leads
Amazon Nova Lite: 26.0 (#254), Qwen3.5 122B-A10B: 38.2 (#138)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| Vectara Hallucination Rate | 6.1% | 11.2% |
| LMArena Expert | 1201 | 1431 |
| Humanity's Last Exam | 3.6% | — |
| MMLU-Pro | 60% | — |
| GPQA (HELM) | 39.7% | — |
| MMLU | 77% | — |
Multimodal Qwen3.5 122B-A10B leads
Amazon Nova Lite: 25.0 (#118), Qwen3.5 122B-A10B: 39.0 (#52)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Vision | 990 | 1245 |
Multilingual Qwen3.5 122B-A10B leads
Amazon Nova Lite: 38.2 (#227), Qwen3.5 122B-A10B: 51.8 (#101)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Non-English | 1208 | 1400 |
| LMArena Chinese | 1225 | 1462 |
| LMArena French | 1237 | 1442 |
| LMArena German | 1229 | 1425 |
| LMArena Japanese | 1153 | 1364 |
| LMArena Korean | 1153 | 1353 |
| LMArena Russian | 1216 | 1400 |
| LMArena Spanish | 1225 | 1423 |
Instruction Following Qwen3.5 122B-A10B leads
Amazon Nova Lite: 59.6 (#250), Qwen3.5 122B-A10B: 74.0 (#109)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Instruction Following | 1205 | 1399 |
| LiveBench Instruction Following | 54.1% | — |
| IFEval | 77.6% | — |
Long Context Qwen3.5 122B-A10B leads
Amazon Nova Lite: 37.4 (#212), Qwen3.5 122B-A10B: 43.0 (#105)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Longer Query | 1234 | 1410 |
Writing & Preference Qwen3.5 122B-A10B leads
Amazon Nova Lite: 42.3 (#232), Qwen3.5 122B-A10B: 60.2 (#98)
| Benchmark | Amazon Nova Lite | Qwen3.5 122B-A10B |
|---|---|---|
| LMArena Text | 1229 | 1417 |
| LMArena Creative Writing | 1197 | 1369 |
| LMArena Multi-Turn | 1199 | 1416 |
| WildBench | 75% | — |
| LiveBench Language | 25.9% | — |
Frequently asked questions
Is Amazon Nova Lite better than Qwen3.5 122B-A10B?
Qwen3.5 122B-A10B is the stronger model overall, scoring 41.6 to 31.4 on the Noometry Index. Amazon Nova Lite costs 10× 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, Amazon Nova Lite or Qwen3.5 122B-A10B?
Amazon Nova Lite is cheaper. It lists at $0.06 per million input tokens and $0.24 per million output tokens; Qwen3.5 122B-A10B lists at $0.40 and $3.20.
Is Amazon Nova Lite or Qwen3.5 122B-A10B better for coding?
Qwen3.5 122B-A10B scores higher on coding benchmarks: 38.8 versus 32.0 in the Noometry coding category.
Which has the bigger context window?
Amazon Nova Lite does, with 300K tokens against 262K.
How many benchmarks do Amazon Nova Lite and Qwen3.5 122B-A10B share?
19 benchmarks have published results for both models. Amazon Nova Lite has 34 scored results on Noometry and Qwen3.5 122B-A10B has 27.