Model comparison
DeepSeek-V2.5 (Sep 2024) vs Tulu 3 (Tülu 3) 70B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 32.6 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-V2.5 (Sep 2024) scores higher in 7 categories and Tulu 3 (Tülu 3) 70B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 33.3 to 13.3.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B | |
|---|---|---|
| Provider | DeepSeek | Allen Institute for AI (Ai2) |
| Noometry Index | 37.0 | 32.6 |
| Released | 2024-09-06 | 2024-11-21 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Tulu 3 (Tülu 3) 70B leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), Tulu 3 (Tülu 3) 70B: 35.4 (#211)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Coding | 1310 | 1236 |
| Aider Polyglot | 17.8% | — |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), Tulu 3 (Tülu 3) 70B: 22.9 (#172)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1220 |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), Tulu 3 (Tülu 3) 70B: 13.3 (#297)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Math | 1289 | 1242 |
| OTIS Mock AIME 2024-2025 | — | 4.4% |
| MATH Level 5 | — | 42.7% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), Tulu 3 (Tülu 3) 70B: 25.1 (#257)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| GPQA Diamond | — | 46.3% |
| LMArena Expert | 1266 | — |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), Tulu 3 (Tülu 3) 70B: 40.1 (#217)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Non-English | 1273 | 1236 |
| LMArena Chinese | 1317 | 1249 |
| LMArena Russian | 1289 | 1246 |
| LMArena French | 1288 | — |
| LMArena German | 1258 | — |
| LMArena Japanese | 1228 | — |
| LMArena Korean | 1209 | — |
| LMArena Spanish | 1248 | — |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), Tulu 3 (Tülu 3) 70B: 65.0 (#222)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1233 |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), Tulu 3 (Tülu 3) 70B: 37.1 (#217)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Longer Query | 1301 | 1224 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), Tulu 3 (Tülu 3) 70B: 45.8 (#218)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Text | 1294 | 1256 |
| LMArena Creative Writing | 1285 | 1231 |
| LMArena Multi-Turn | 1297 | 1252 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Tulu 3 (Tülu 3) 70B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 32.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Tulu 3 (Tülu 3) 70B better for coding?
Tulu 3 (Tülu 3) 70B scores higher on coding benchmarks: 35.4 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Tulu 3 (Tülu 3) 70B share?
11 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Tulu 3 (Tülu 3) 70B has 14.