Model comparison
DeepSeek-V2.5 (Sep 2024) vs Olmo 2 0325 32b Instruct
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 32.7 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 Olmo 2 0325 32b Instruct in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V2.5 (Sep 2024) leads 34.8 to 19.5.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct | |
|---|---|---|
| Provider | DeepSeek | Allen Institute for AI (Ai2) |
| Noometry Index | 37.6 | 32.7 |
| Released | 2024-09-06 | — |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 16 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Olmo 2 0325 32b Instruct leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Olmo 2 0325 32b Instruct: 35.2 (#227)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Coding | 1309 | 1210 |
| 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): 25.6 (#145), Olmo 2 0325 32b Instruct: 23.6 (#175)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1208 |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Olmo 2 0325 32b Instruct: 26.8 (#255)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Math | 1288 | 1208 |
| Omni-MATH | — | 16.1% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Olmo 2 0325 32b Instruct: 19.5 (#279)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| MMLU-Pro | — | 41.4% |
| GPQA (HELM) | — | 28.7% |
| LMArena Expert | 1266 | — |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Olmo 2 0325 32b Instruct: 34.8 (#248)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Non-English | 1273 | 1160 |
| LMArena Chinese | 1318 | 1192 |
| LMArena Russian | 1289 | 1187 |
| LMArena French | 1289 | — |
| 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): 67.5 (#194), Olmo 2 0325 32b Instruct: 61.5 (#244)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Instruction Following | 1280 | 1186 |
| IFEval | — | 78% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Olmo 2 0325 32b Instruct: 36.2 (#234)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Longer Query | 1301 | 1194 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Olmo 2 0325 32b Instruct: 42.1 (#236)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Olmo 2 0325 32b Instruct |
|---|---|---|
| LMArena Text | 1294 | 1218 |
| LMArena Creative Writing | 1285 | 1199 |
| LMArena Multi-Turn | 1297 | 1221 |
| WildBench | — | 73.4% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Olmo 2 0325 32b Instruct?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 32.7 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Olmo 2 0325 32b Instruct better for coding?
Olmo 2 0325 32b Instruct scores higher on coding benchmarks: 35.2 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Olmo 2 0325 32b Instruct share?
11 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Olmo 2 0325 32b Instruct has 16.