Model comparison
DeepSeek-V2.5 (Sep 2024) vs GPT-4.1 mini
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 33.6 on the Noometry Index.
Last verified . 19 shared benchmarks.
DeepSeekSummary
- They share 19 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 6 categories and GPT-4.1 mini in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V2.5 (Sep 2024) leads 25.6 to 10.8.
- The biggest single-benchmark swing is Aider Polyglot: 17.8% for DeepSeek-V2.5 (Sep 2024) and 32.4% for GPT-4.1 mini.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 37.6 | 33.6 |
| Released | 2024-09-06 | 2025-04-14 |
| Weights | Open | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 33K |
| Input $ / M tokens | — | $0.40 |
| Output $ / M tokens | — | $1.60 |
| Results tracked | 22 | 47 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), GPT-4.1 mini: 30.6 (#293)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| Aider Polyglot | 17.8% | 32.4% |
| BigCodeBench Instruct | 48.6% | 48.9% |
| LMArena Coding | 1309 | 1367 |
| SWE-bench Verified (bash only) | — | 23.9% |
| SciCode | — | 40.4% |
| WeirdML | — | 37.6% |
| BigCodeBench Complete | 53.2% | — |
| CadEval | — | 16% |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, GPT-4.1 mini: 33.3 (#55)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 50.5% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), GPT-4.1 mini: 10.8 (#340)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1349 |
| ARC-AGI-2 | — | 0% |
| Kagi LLM Benchmark | — | 48.6% |
| ARC-AGI-1 | — | 3.5% |
| CritPt | — | 0% |
| Chess Puzzles | — | 7% |
| Mystery Game Puzzles | — | 7% |
| DTBench | — | 68.8% |
| LMCA | — | 21.1% |
| Epoch Capabilities Index | — | 135.01 |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), GPT-4.1 mini: 24.1 (#270)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Math | 1288 | 1343 |
| FrontierMath (Tiers 1-3) | — | 6.7% |
| OTIS Mock AIME 2024-2025 | — | 44.7% |
| Omni-MATH | — | 49.1% |
| MATH Level 5 | — | 87.3% |
| FrontierMath (Feb 2025 set) | — | 4.5% |
Knowledge Too close to call
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), GPT-4.1 mini: 34.7 (#194)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Expert | 1266 | 1338 |
| GPQA Diamond | — | 65.8% |
| SimpleQA Verified | — | 12.7% |
| MMLU-Pro | — | 78.3% |
| GPQA (HELM) | — | 61.4% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, GPT-4.1 mini: 35.8 (#82)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Vision | — | 1181 |
Multilingual GPT-4.1 mini leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), GPT-4.1 mini: 45.7 (#166)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Non-English | 1273 | 1318 |
| LMArena Chinese | 1318 | 1329 |
| LMArena French | 1289 | 1358 |
| LMArena German | 1258 | 1351 |
| LMArena Japanese | 1228 | 1290 |
| LMArena Korean | 1209 | 1298 |
| LMArena Russian | 1289 | 1324 |
| LMArena Spanish | 1248 | 1319 |
Instruction Following GPT-4.1 mini leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), GPT-4.1 mini: 73.7 (#118)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Instruction Following | 1280 | 1333 |
| IFEval | — | 90.4% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), GPT-4.1 mini: 31.8 (#275)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Longer Query | 1301 | 1344 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), GPT-4.1 mini: 48.6 (#199)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | GPT-4.1 mini |
|---|---|---|
| LMArena Text | 1294 | 1340 |
| LMArena Creative Writing | 1285 | 1300 |
| LMArena Multi-Turn | 1297 | 1354 |
| EQ-Bench Creative Writing | — | 1147 |
| WildBench | — | 83.8% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than GPT-4.1 mini?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 33.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or GPT-4.1 mini better for coding?
DeepSeek-V2.5 (Sep 2024) scores higher on coding benchmarks: 31.7 versus 30.6 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and GPT-4.1 mini share?
19 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and GPT-4.1 mini has 47.