Model comparison
Kimi K3 vs Llama 3-70B
Kimi K3 is the stronger model overall, scoring 59.5 to 28.5 on the Noometry Index.
Last verified . 22 shared benchmarks.
MetaSummary
- They share 22 benchmarks with published results for both. Kimi K3 scores higher in 9 categories and Llama 3-70B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.0 to 12.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 97.2% for Kimi K3 and 4.3% for Llama 3-70B.
Side by side
| Kimi K3 | Llama 3-70B | |
|---|---|---|
| Provider | Moonshot AI | Meta |
| Noometry Index | 59.5 | 28.5 |
| Released | 2026-07-16 | 2024-04-18 |
| Weights | Open | Open |
| Context window | 1.05M | — |
| Max output | 1.05M | — |
| Input $ / M tokens | $3 | — |
| Output $ / M tokens | $15 | — |
| Results tracked | 53 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Kimi K3: 60.6 (#8), Llama 3-70B: 35.3 (#215)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Coding | 1509 | 1206 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| LMArena WebDev | 1659 | — |
| FrontierSWE | 25.9% | — |
| SciCode | 59.5% | — |
| WeirdML | 82.6% | — |
| BigCodeBench Instruct | — | 43.6% |
| BigCodeBench Complete | — | 54.5% |
| ALE-Bench | 1,524 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 69% |
Agentic & Tool Use Kimi K3 leads
Kimi K3: 41.8 (#19), Llama 3-70B: 20.9 (#132)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| APEX-Agents | 50.6% | — |
| τ²-bench Banking | 37.1% | — |
| Cybench | — | 5% |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |
| Vending-Bench 2 | 5,165 | — |
Reasoning Kimi K3 leads
Kimi K3: 63.0 (#15), Llama 3-70B: 17.4 (#284)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Hard Prompts | 1495 | 1195 |
| DTBench | 91.2% | 54.2% |
| Epoch Capabilities Index | 157.68 | 122.91 |
| ForecastBench | 61.1 | 57.1 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | 60.7% | — |
| Kagi LLM Benchmark | — | 35.1% |
| NYT Connections (extended) | 93.6% | — |
| ARC-AGI-1 | 94.5% | — |
| CritPt | 23.4% | — |
| Chess Puzzles | 39% | — |
| Mystery Game Puzzles | 26% | — |
| LMCA | 52.7% | — |
| Surface Evolver Bench | 95% | — |
| WinoGrande | — | 83.5% |
Math Kimi K3 leads
Kimi K3: 74.0 (#13), Llama 3-70B: 12.0 (#299)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 97.2% | 4.3% |
| LMArena Math | 1493 | 1218 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| ProofBench | 87% | — |
| MATH Level 5 | — | 22.6% |
Knowledge Kimi K3 leads
Kimi K3: 63.2 (#19), Llama 3-70B: 20.8 (#272)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| GPQA Diamond | 93.1% | 40.6% |
| LMArena Expert | 1520 | 1149 |
| SimpleQA Verified | 50.6% | — |
| MMLU | — | 79.3% |
Multimodal Not comparable
Kimi K3: 38.4 (#58), Llama 3-70B: —
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |
Multilingual Kimi K3 leads
Kimi K3: 56.5 (#20), Llama 3-70B: 33.8 (#246)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Non-English | 1466 | 1142 |
| LMArena Chinese | 1528 | 1114 |
| LMArena French | 1488 | 1231 |
| LMArena German | 1474 | 1169 |
| LMArena Japanese | 1492 | 1017 |
| LMArena Korean | 1465 | 1017 |
| LMArena Russian | 1481 | 1159 |
| LMArena Spanish | 1478 | 1241 |
Instruction Following Kimi K3 leads
Kimi K3: 77.8 (#13), Llama 3-70B: 62.8 (#233)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Instruction Following | 1481 | 1194 |
Long Context Kimi K3 leads
Kimi K3: 45.8 (#27), Llama 3-70B: 35.6 (#235)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Longer Query | 1494 | 1174 |
Writing & Preference Kimi K3 leads
Kimi K3: 76.7 (#4), Llama 3-70B: 43.0 (#225)
| Benchmark | Kimi K3 | Llama 3-70B |
|---|---|---|
| LMArena Text | 1476 | 1221 |
| LMArena Creative Writing | 1452 | 1210 |
| LMArena Multi-Turn | 1488 | 1223 |
| EQ-Bench Creative Writing | 2082 | — |
| EQ-Bench 4 | 1339 | — |
Frequently asked questions
Is Kimi K3 better than Llama 3-70B?
Kimi K3 is the stronger model overall, scoring 59.5 to 28.5 on the Noometry Index.
Is Kimi K3 or Llama 3-70B better for coding?
Kimi K3 scores higher on coding benchmarks: 60.6 versus 35.3 in the Noometry coding category.
How many benchmarks do Kimi K3 and Llama 3-70B share?
22 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Llama 3-70B has 31.