Model comparison
DeepSeek V4 Pro vs Granite 3.1 2b Instruct
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 33.2 on the Noometry Index.
Last verified . 12 shared benchmarks.
DeepSeekSummary
- They share 12 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and Granite 3.1 2b Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 22.0.
Side by side
| DeepSeek V4 Pro | Granite 3.1 2b Instruct | |
|---|---|---|
| Provider | DeepSeek | IBM |
| Noometry Index | 54.3 | 33.2 |
| Released | 2026-04-24 | — |
| Weights | Open | Open |
| Context window | 1M | — |
| Max output | 393K | — |
| Input $ / M tokens | $0.66 | — |
| Output $ / M tokens | $1.98 | — |
| Results tracked | 48 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Granite 3.1 2b Instruct: 33.4 (#257)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Coding | 1470 | 1149 |
| 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: 32.8 (#58), Granite 3.1 2b Instruct: —
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Granite 3.1 2b Instruct: 22.0 (#209)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Hard Prompts | 1461 | 1138 |
| 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% | — |
| Epoch Capabilities Index | 155.31 | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Granite 3.1 2b Instruct: 33.1 (#206)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Math | 1455 | 1159 |
| 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 DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Granite 3.1 2b Instruct: 30.8 (#224)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Expert | 1464 | 1131 |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Granite 3.1 2b Instruct: 29.1 (#269)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Non-English | 1439 | 1068 |
| LMArena Chinese | 1486 | 1139 |
| LMArena Russian | 1453 | 1063 |
| LMArena French | 1472 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1445 | — |
| LMArena Korean | 1447 | — |
| LMArena Spanish | 1458 | — |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), Granite 3.1 2b Instruct: 57.7 (#264)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Instruction Following | 1448 | 1116 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Granite 3.1 2b Instruct: 35.0 (#244)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Longer Query | 1458 | 1155 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Granite 3.1 2b Instruct: 34.1 (#274)
| Benchmark | DeepSeek V4 Pro | Granite 3.1 2b Instruct |
|---|---|---|
| LMArena Text | 1451 | 1127 |
| LMArena Creative Writing | 1446 | 1116 |
| LMArena Multi-Turn | 1467 | 1099 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Granite 3.1 2b Instruct?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 33.2 on the Noometry Index.
Is DeepSeek V4 Pro or Granite 3.1 2b Instruct better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 33.4 in the Noometry coding category.
How many benchmarks do DeepSeek V4 Pro and Granite 3.1 2b Instruct share?
12 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Granite 3.1 2b Instruct has 12.