Model comparison
Deepseek Coder v2 vs Gemini 2.5 Flash
Gemini 2.5 Flash is the stronger model overall, scoring 39.0 to 35.3 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
GoogleSummary
- They share 17 benchmarks with published results for both. Deepseek Coder v2 scores higher in 2 categories and Gemini 2.5 Flash in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Gemini 2.5 Flash leads 52.5 to 36.4.
- Deepseek Coder v2 has downloadable open weights; the other is API-only.
Side by side
| Deepseek Coder v2 | Gemini 2.5 Flash | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 35.3 | 39.0 |
| Released | 2024-06-17 | 2025-04-17 |
| Weights | Open | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.30 |
| Output $ / M tokens | — | $2.50 |
| Results tracked | 24 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 37.5 (#178), Gemini 2.5 Flash: 35.5 (#210)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Coding | 1251 | 1424 |
| SWE-bench Verified (bash only) | — | 28.7% |
| Aider Polyglot | — | 55.1% |
| WeirdML | — | 41.9% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| ALE-Bench | — | 661.88 |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, Gemini 2.5 Flash: 30.6 (#67)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| Terminal-Bench | — | 17.1% |
| Berkeley Function Calling Leaderboard | — | 56.2% |
| TheAgentCompany | — | 41.1% |
| BALROG | — | 33.5% |
| Vending-Bench 2 | — | 548.84 |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 22.6 (#175), Gemini 2.5 Flash: 18.2 (#271)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1422 |
| ARC-AGI-2 | — | 2.5% |
| SimpleBench | — | 41.2% |
| Kagi LLM Benchmark | — | 56.8% |
| ARC-AGI-1 | — | 33.3% |
| CritPt | — | 1.1% |
| EnigmaEval | — | 2.7% |
| DTBench | — | 76.5% |
| LMCA | — | 27.5% |
| Epoch Capabilities Index | — | 143.04 |
| ForecastBench | — | 60.6 |
| WinoGrande | 83.7% | — |
Math Gemini 2.5 Flash leads
Deepseek Coder v2: 32.3 (#188), Gemini 2.5 Flash: 38.1 (#87)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Math | 1241 | 1415 |
| OTIS Mock AIME 2024-2025 | — | 73.1% |
| Omni-MATH | — | 38.5% |
| FrontierMath (Feb 2025 set) | — | 4.8% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
| GSM8K | 94.5% | — |
Knowledge Gemini 2.5 Flash leads
Deepseek Coder v2: 31.7 (#209), Gemini 2.5 Flash: 36.1 (#162)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Expert | 1181 | 1426 |
| Humanity's Last Exam | — | 12.1% |
| MMLU-Pro | — | 63.9% |
| Confabulations | — | 16.8% |
| Vectara Hallucination Rate | — | 7.8% |
| GPQA (HELM) | — | 39% |
| ARC (AI2) Challenge | 64.3% | — |
Multimodal Not comparable
Deepseek Coder v2: —, Gemini 2.5 Flash: 41.2 (#31)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Vision | — | 1254 |
| GeoBench | — | 76% |
| VPCT | — | 46.2% |
| SpatialViz-Bench | — | 36.9% |
Multilingual Gemini 2.5 Flash leads
Deepseek Coder v2: 36.4 (#235), Gemini 2.5 Flash: 52.5 (#83)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Non-English | 1182 | 1410 |
| LMArena Chinese | 1201 | 1450 |
| LMArena French | 1184 | 1433 |
| LMArena German | 1164 | 1419 |
| LMArena Japanese | 1126 | 1403 |
| LMArena Korean | 1104 | 1385 |
| LMArena Russian | 1188 | 1415 |
| LMArena Spanish | 1153 | 1421 |
Instruction Following Gemini 2.5 Flash leads
Deepseek Coder v2: 61.9 (#237), Gemini 2.5 Flash: 75.9 (#50)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Instruction Following | 1180 | 1405 |
| IFEval | — | 89.8% |
Long Context Gemini 2.5 Flash leads
Deepseek Coder v2: 37.0 (#219), Gemini 2.5 Flash: 47.5 (#16)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Longer Query | 1219 | 1419 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference Gemini 2.5 Flash leads
Deepseek Coder v2: 38.4 (#249), Gemini 2.5 Flash: 54.0 (#153)
| Benchmark | Deepseek Coder v2 | Gemini 2.5 Flash |
|---|---|---|
| LMArena Text | 1191 | 1417 |
| LMArena Creative Writing | 1120 | 1400 |
| LMArena Multi-Turn | 1176 | 1408 |
| Short-Story Creative Writing | — | 76.5% |
| EQ-Bench Creative Writing | — | 1137 |
| WildBench | — | 81.7% |
Frequently asked questions
Is Deepseek Coder v2 better than Gemini 2.5 Flash?
Gemini 2.5 Flash is the stronger model overall, scoring 39.0 to 35.3 on the Noometry Index.
Is Deepseek Coder v2 or Gemini 2.5 Flash better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 37.5 versus 35.5 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Gemini 2.5 Flash share?
17 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Gemini 2.5 Flash has 54.