Model comparison
Amazon Nova Micro vs DeepSeek-V3
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 30.4 on the Noometry Index. Amazon Nova Micro costs 6.6× less per token, which makes it the better buy when DeepSeek-V3's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
DeepSeekSummary
- They share 31 benchmarks with published results for both. Amazon Nova Micro scores higher in 1 category and DeepSeek-V3 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-V3 leads 57.4 to 39.5.
- The biggest single-benchmark swing is LiveBench Coding: 20.2% for Amazon Nova Micro and 70.9% for DeepSeek-V3.
- Amazon Nova Micro is cheaper at $0.035 / $0.14 per million input/output tokens, against $0.24 / $0.90 for DeepSeek-V3.
- DeepSeek-V3 accepts more context: 164K tokens versus 128K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| Amazon Nova Micro | DeepSeek-V3 | |
|---|---|---|
| Provider | Amazon | DeepSeek |
| Noometry Index | 30.4 | 39.5 |
| Released | 2024-12-03 | 2024-12-26 |
| Weights | Proprietary | Open |
| Context window | 128K | 164K |
| Max output | 10K | 164K |
| Input $ / M tokens | $0.035 | $0.24 |
| Output $ / M tokens | $0.14 | $0.90 |
| Results tracked | 32 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3 leads
Amazon Nova Micro: 30.5 (#295), DeepSeek-V3: 42.3 (#106)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LiveBench Coding | 20.2% | 70.9% |
| LMArena Coding | 1218 | 1368 |
| Aider Polyglot | — | 55.1% |
| SciCode | — | 35.8% |
| WeirdML | — | 36.1% |
| BigCodeBench Instruct | — | 50% |
| BigCodeBench Complete | — | 62.2% |
| HumanEval+ | — | 86.6% |
| MBPP+ | — | 73% |
Agentic & Tool Use Not comparable
Amazon Nova Micro: 22.1 (#132), DeepSeek-V3: —
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 22.3% | — |
| METR Time Horizons | — | 49.6% |
Reasoning DeepSeek-V3 leads
Amazon Nova Micro: 17.4 (#294), DeepSeek-V3: 20.5 (#236)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LiveBench Reasoning | 25.1% | 65.8% |
| LMArena Hard Prompts | 1191 | 1365 |
| LiveBench Data Analysis | 34% | 60.9% |
| LiveBench | 29.6% | 66.9% |
| SimpleBench | — | 27.2% |
| Kagi LLM Benchmark | — | 52.3% |
| CritPt | — | 0% |
| DTBench | — | 64.8% |
| LMCA | — | 15.5% |
| BIG-Bench Hard | — | 87.5% |
| Epoch Capabilities Index | — | 135.94 |
| ForecastBench | — | 59.1 |
| HellaSwag | — | 88.9% |
| PIQA | — | 84.7% |
| WinoGrande | — | 85.2% |
Math DeepSeek-V3 leads
Amazon Nova Micro: 26.9 (#254), DeepSeek-V3: 32.1 (#219)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| Omni-MATH | 21.4% | 40.3% |
| LiveBench Math | 34.5% | 73.5% |
| LMArena Math | 1206 | 1373 |
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| MATH Level 5 | — | 75.5% |
| FrontierMath (Feb 2025 set) | — | 1.7% |
Knowledge DeepSeek-V3 leads
Amazon Nova Micro: 29.6 (#237), DeepSeek-V3: 37.5 (#155)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| MMLU-Pro | 51.1% | 72.3% |
| Vectara Hallucination Rate | 5.5% | 6.1% |
| GPQA (HELM) | 38.3% | 53.8% |
| LMArena Expert | 1184 | 1351 |
| MMLU | 70.8% | 87.2% |
| GPQA Diamond | — | 67.6% |
| Confabulations | — | 26.1% |
| ARC (AI2) Challenge | — | 95.3% |
| TriviaQA | — | 82.9% |
Multilingual DeepSeek-V3 leads
Amazon Nova Micro: 36.5 (#239), DeepSeek-V3: 48.5 (#143)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LMArena Non-English | 1186 | 1358 |
| LMArena Chinese | 1209 | 1391 |
| LMArena French | 1238 | 1385 |
| LMArena German | 1192 | 1374 |
| LMArena Japanese | 1154 | 1333 |
| LMArena Korean | 1150 | 1319 |
| LMArena Russian | 1185 | 1373 |
| LMArena Spanish | 1225 | 1358 |
Instruction Following DeepSeek-V3 leads
Amazon Nova Micro: 56.3 (#272), DeepSeek-V3: 72.8 (#130)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LiveBench Instruction Following | 48% | 81.5% |
| IFEval | 76% | 83.2% |
| LMArena Instruction Following | 1174 | 1345 |
Long Context Amazon Nova Micro leads
Amazon Nova Micro: 36.5 (#229), DeepSeek-V3: 34.0 (#253)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LMArena Longer Query | 1205 | 1352 |
| Fiction.LiveBench | — | 50% |
Writing & Preference DeepSeek-V3 leads
Amazon Nova Micro: 39.5 (#247), DeepSeek-V3: 57.4 (#130)
| Benchmark | Amazon Nova Micro | DeepSeek-V3 |
|---|---|---|
| LMArena Text | 1208 | 1375 |
| LMArena Creative Writing | 1172 | 1364 |
| WildBench | 74.3% | 83% |
| LMArena Multi-Turn | 1178 | 1389 |
| LiveBench Language | 15.8% | 49.1% |
| Short-Story Creative Writing | — | 77% |
| EQ-Bench Creative Writing | — | 1472 |
Frequently asked questions
Is Amazon Nova Micro better than DeepSeek-V3?
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 30.4 on the Noometry Index. Amazon Nova Micro costs 6.6× less per token, which makes it the better buy when DeepSeek-V3's lead doesn't matter for your workload.
Which is cheaper, Amazon Nova Micro or DeepSeek-V3?
Amazon Nova Micro is cheaper. It lists at $0.035 per million input tokens and $0.14 per million output tokens; DeepSeek-V3 lists at $0.24 and $0.90.
Is Amazon Nova Micro or DeepSeek-V3 better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 30.5 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3 does, with 164K tokens against 128K.
How many benchmarks do Amazon Nova Micro and DeepSeek-V3 share?
31 benchmarks have published results for both models. Amazon Nova Micro has 32 scored results on Noometry and DeepSeek-V3 has 60.