Model comparison
DeepSeek-V2.5 (Sep 2024) vs Qwen3 14B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 35.5 on the Noometry Index.
Last verified . 0 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- The widest gap is in reasoning, where DeepSeek-V2.5 (Sep 2024) leads 25.6 to 18.5.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Qwen3 14B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 37.6 | 35.5 |
| Released | 2024-09-06 | 2025-04 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $0.35 |
| Output $ / M tokens | — | $1.40 |
| Results tracked | 22 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 14B leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Qwen3 14B: 37.3 (#195)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| Aider Polyglot | 17.8% | — |
| SciCode | — | 31.6% |
| BigCodeBench Instruct | 48.6% | — |
| LMArena Coding | 1309 | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Qwen3 14B: 29.6 (#83)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 41% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Qwen3 14B: 18.5 (#280)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| Kagi LLM Benchmark | — | 49.1% |
| CritPt | — | 0% |
| Chess Puzzles | — | 4% |
| LMArena Hard Prompts | 1289 | — |
| DTBench | — | 64% |
| LMCA | — | 18.2% |
| Epoch Capabilities Index | — | 138.23 |
Math Qwen3 14B leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Qwen3 14B: 38.6 (#133)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 66.4% |
| LMArena Math | 1288 | — |
Knowledge Qwen3 14B leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Qwen3 14B: 39.3 (#134)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| GPQA Diamond | — | 63.8% |
| Vectara Hallucination Rate | — | 5.4% |
| LMArena Expert | 1266 | — |
Multilingual Not comparable
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Qwen3 14B: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| LMArena Non-English | 1273 | — |
| LMArena Chinese | 1318 | — |
| LMArena French | 1289 | — |
| LMArena German | 1258 | — |
| LMArena Japanese | 1228 | — |
| LMArena Korean | 1209 | — |
| LMArena Russian | 1289 | — |
| LMArena Spanish | 1248 | — |
Instruction Following Not comparable
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Qwen3 14B: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| LMArena Instruction Following | 1280 | — |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Qwen3 14B: 38.1 (#204)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| Fiction.LiveBench | — | 62.5% |
| LMArena Longer Query | 1301 | — |
Writing & Preference Not comparable
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Qwen3 14B: —
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 14B |
|---|---|---|
| LMArena Text | 1294 | — |
| LMArena Creative Writing | 1285 | — |
| LMArena Multi-Turn | 1297 | — |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Qwen3 14B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 35.5 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Qwen3 14B better for coding?
Qwen3 14B scores higher on coding benchmarks: 37.3 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Qwen3 14B share?
0 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Qwen3 14B has 12.