Model comparison
DeepSeek V4 Pro vs Dolly 2.0-12b
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 24.9 on the Noometry Index.
Last verified . 10 shared benchmarks.
DeepSeek
DatabricksSummary
- They share 10 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 6 categories and Dolly 2.0-12b in 0 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek V4 Pro leads 65.6 to 15.3.
Side by side
| DeepSeek V4 Pro | Dolly 2.0-12b | |
|---|---|---|
| Provider | DeepSeek | Databricks |
| Noometry Index | 54.2 | 24.9 |
| Released | 2026-04-24 | 2023-04-11 |
| Weights | Open | Open |
| Context window | 1M | — |
| Max output | 393K | — |
| Input $ / M tokens | $0.43 | — |
| Output $ / M tokens | $0.87 | — |
| Results tracked | 48 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.2 (#31), Dolly 2.0-12b: 23.0 (#327)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Coding | 1470 | 776 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| WeirdML | 66.2% | — |
| ALE-Bench | 1,403 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.7 (#54), Dolly 2.0-12b: —
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.4 (#22), Dolly 2.0-12b: 14.6 (#315)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Hard Prompts | 1461 | 804 |
| Epoch Capabilities Index | 155.39 | 88.94 |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| CritPt | 18% | — |
| Chess Puzzles | 47% | — |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
| HellaSwag | — | 70.8% |
| PIQA | — | 75.4% |
| WinoGrande | — | 61.8% |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.6 (#24), Dolly 2.0-12b: 25.0 (#250)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Math | 1455 | 871 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
Knowledge Not comparable
DeepSeek V4 Pro: 59.5 (#28), Dolly 2.0-12b: —
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| LMArena Expert | 1463 | — |
| ARC (AI2) Challenge | — | 39.6% |
| BoolQ | — | 56.3% |
| MMLU | — | 26.2% |
| OpenBookQA | — | 39.2% |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.5 (#44), Dolly 2.0-12b: 17.4 (#291)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Non-English | 1438 | 836 |
| LMArena Chinese | 1486 | 836 |
| LMArena French | 1473 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1448 | — |
| LMArena Korean | 1452 | — |
| LMArena Russian | 1452 | — |
| LMArena Spanish | 1457 | — |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.3 (#44), Dolly 2.0-12b: 38.9 (#299)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Instruction Following | 1448 | 814 |
Long Context Not comparable
DeepSeek V4 Pro: 45.0 (#49), Dolly 2.0-12b: —
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| CL-bench Life | 13.5% | — |
| LMArena Longer Query | 1459 | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.6 (#45), Dolly 2.0-12b: 15.3 (#306)
| Benchmark | DeepSeek V4 Pro | Dolly 2.0-12b |
|---|---|---|
| LMArena Text | 1451 | 851 |
| LMArena Creative Writing | 1446 | 864 |
| LMArena Multi-Turn | 1467 | 740 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Dolly 2.0-12b?
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 24.9 on the Noometry Index.
Is DeepSeek V4 Pro or Dolly 2.0-12b better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.2 versus 23.0 in the Noometry coding category.
How many benchmarks do DeepSeek V4 Pro and Dolly 2.0-12b share?
10 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Dolly 2.0-12b has 17.