Model comparison
Deepseek Coder v2 vs Tulu 3 (Tülu 3) 70B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 33.0 on the Noometry Index.
Last verified . 11 shared benchmarks.
DeepSeek
Allen Institute for AI (Ai2)Summary
- They share 11 benchmarks with published results for both. Deepseek Coder v2 scores higher in 3 categories and Tulu 3 (Tülu 3) 70B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Deepseek Coder v2 leads 34.9 to 14.2.
Side by side
| Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B | |
|---|---|---|
| Provider | DeepSeek | Allen Institute for AI (Ai2) |
| Noometry Index | 35.9 | 33.0 |
| Released | 2024-06-17 | 2024-11-21 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 38.1 (#183), Tulu 3 (Tülu 3) 70B: 36.0 (#214)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Coding | 1251 | 1235 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Too close to call
Deepseek Coder v2: 23.6 (#176), Tulu 3 (Tülu 3) 70B: 23.9 (#169)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1220 |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Tulu 3 (Tülu 3) 70B: 14.2 (#303)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Math | 1241 | 1242 |
| OTIS Mock AIME 2024-2025 | — | 4.4% |
| MATH Level 5 | — | 42.7% |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Tulu 3 (Tülu 3) 70B: 25.0 (#264)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| GPQA Diamond | — | 46.3% |
| LMArena Expert | 1181 | — |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual Tulu 3 (Tülu 3) 70B leads
Deepseek Coder v2: 36.3 (#240), Tulu 3 (Tülu 3) 70B: 39.9 (#222)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Non-English | 1182 | 1236 |
| LMArena Chinese | 1201 | 1249 |
| LMArena Russian | 1188 | 1246 |
| LMArena French | 1185 | — |
| LMArena German | 1164 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Tulu 3 (Tülu 3) 70B leads
Deepseek Coder v2: 61.7 (#242), Tulu 3 (Tülu 3) 70B: 64.8 (#227)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1233 |
Long Context Too close to call
Deepseek Coder v2: 37.0 (#224), Tulu 3 (Tülu 3) 70B: 37.1 (#222)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Longer Query | 1219 | 1224 |
Writing & Preference Tulu 3 (Tülu 3) 70B leads
Deepseek Coder v2: 38.2 (#253), Tulu 3 (Tülu 3) 70B: 45.6 (#223)
| Benchmark | Deepseek Coder v2 | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Text | 1191 | 1256 |
| LMArena Creative Writing | 1120 | 1231 |
| LMArena Multi-Turn | 1177 | 1252 |
Frequently asked questions
Is Deepseek Coder v2 better than Tulu 3 (Tülu 3) 70B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 33.0 on the Noometry Index.
Is Deepseek Coder v2 or Tulu 3 (Tülu 3) 70B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 36.0 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Tulu 3 (Tülu 3) 70B share?
11 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Tulu 3 (Tülu 3) 70B has 14.