Model comparison
Llama 3-70B vs Qwen3.5 397B-A17B
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 28.8 on the Noometry Index.
Last verified . 22 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. Llama 3-70B scores higher in 0 categories and Qwen3.5 397B-A17B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.5 397B-A17B leads 46.1 to 12.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 4.3% for Llama 3-70B and 88.9% for Qwen3.5 397B-A17B.
Side by side
| Llama 3-70B | Qwen3.5 397B-A17B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.8 | 46.0 |
| Released | 2024-04-18 | 2026-02-01 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.60 |
| Output $ / M tokens | — | $3.60 |
| Results tracked | 31 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 397B-A17B leads
Llama 3-70B: 35.8 (#218), Qwen3.5 397B-A17B: 42.0 (#114)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Coding | 1206 | 1465 |
| LMArena WebDev | — | 1400 |
| BigCodeBench Instruct | 43.6% | — |
| BigCodeBench Complete | 54.5% | — |
| HumanEval+ | 72% | — |
| MBPP+ | 69% | — |
Agentic & Tool Use Qwen3.5 397B-A17B leads
Llama 3-70B: 21.1 (#139), Qwen3.5 397B-A17B: 33.3 (#53)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| APEX-Agents | — | 24.9% |
| τ²-bench Airline | — | 81.5% |
| τ²-bench Banking | — | 9.8% |
| τ²-bench Retail | — | 84.4% |
| τ²-bench Telecom | — | 97.8% |
| Cybench | 5% | — |
Reasoning Qwen3.5 397B-A17B leads
Llama 3-70B: 18.0 (#288), Qwen3.5 397B-A17B: 34.5 (#70)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| Kagi LLM Benchmark | 35.1% | 73.7% |
| LMArena Hard Prompts | 1195 | 1448 |
| DTBench | 54.2% | 87.5% |
| Epoch Capabilities Index | 122.93 | 146.65 |
| NYT Connections (extended) | — | 58.9% |
| Chess Puzzles | — | 13% |
| Thematic Generalization | — | 65.1% |
| Mystery Game Puzzles | — | 18% |
| LMCA | — | 37.9% |
| ForecastBench | 57.1 | — |
| WinoGrande | 83.5% | — |
Math Qwen3.5 397B-A17B leads
Llama 3-70B: 12.8 (#305), Qwen3.5 397B-A17B: 46.1 (#73)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 88.9% |
| LMArena Math | 1218 | 1454 |
| FrontierMath (Tiers 1-3) | — | 31.2% |
| MATH Level 5 | 22.6% | — |
Knowledge Qwen3.5 397B-A17B leads
Llama 3-70B: 20.8 (#277), Qwen3.5 397B-A17B: 53.3 (#58)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| GPQA Diamond | 40.6% | 86.4% |
| LMArena Expert | 1149 | 1462 |
| MMLU | 79.3% | — |
Multimodal Not comparable
Llama 3-70B: —, Qwen3.5 397B-A17B: 40.7 (#44)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Vision | — | 1263 |
Multilingual Qwen3.5 397B-A17B leads
Llama 3-70B: 33.6 (#251), Qwen3.5 397B-A17B: 53.7 (#59)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Non-English | 1142 | 1430 |
| LMArena Chinese | 1114 | 1500 |
| LMArena French | 1232 | 1461 |
| LMArena German | 1169 | 1447 |
| LMArena Japanese | 1017 | 1426 |
| LMArena Korean | 1017 | 1384 |
| LMArena Russian | 1159 | 1429 |
| LMArena Spanish | 1241 | 1441 |
Instruction Following Qwen3.5 397B-A17B leads
Llama 3-70B: 62.5 (#238), Qwen3.5 397B-A17B: 75.0 (#77)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Instruction Following | 1194 | 1424 |
Long Context Qwen3.5 397B-A17B leads
Llama 3-70B: 35.6 (#240), Qwen3.5 397B-A17B: 44.1 (#74)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Longer Query | 1174 | 1442 |
Writing & Preference Qwen3.5 397B-A17B leads
Llama 3-70B: 42.8 (#231), Qwen3.5 397B-A17B: 62.3 (#79)
| Benchmark | Llama 3-70B | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Text | 1221 | 1438 |
| LMArena Creative Writing | 1210 | 1401 |
| LMArena Multi-Turn | 1223 | 1446 |
| EQ-Bench Creative Writing | — | 1478 |
Frequently asked questions
Is Llama 3-70B better than Qwen3.5 397B-A17B?
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 28.8 on the Noometry Index.
Is Llama 3-70B or Qwen3.5 397B-A17B better for coding?
Qwen3.5 397B-A17B scores higher on coding benchmarks: 42.0 versus 35.8 in the Noometry coding category.
How many benchmarks do Llama 3-70B and Qwen3.5 397B-A17B share?
22 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and Qwen3.5 397B-A17B has 36.