Model comparison
DeepSeek LLM 67B vs DeepSeek V4 Pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 24.9 on the Noometry Index.
Last verified . 14 shared benchmarks.
DeepSeekSummary
- They share 14 benchmarks with published results for both. DeepSeek LLM 67B scores higher in 0 categories and DeepSeek V4 Pro in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.8 to 8.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 0.8% for DeepSeek LLM 67B and 98.6% for DeepSeek V4 Pro.
Side by side
| DeepSeek LLM 67B | DeepSeek V4 Pro | |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Noometry Index | 24.9 | 54.3 |
| Released | 2023-11-29 | 2026-04-24 |
| Weights | Open | Open |
| Context window | — | 1M |
| Max output | — | 393K |
| Input $ / M tokens | — | $0.66 |
| Output $ / M tokens | — | $1.98 |
| Results tracked | 15 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek LLM 67B: 31.9 (#278), DeepSeek V4 Pro: 52.4 (#34)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| LMArena Coding | 1096 | 1470 |
| 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 LLM 67B: —, DeepSeek V4 Pro: 32.8 (#58)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| APEX-Agents | — | 47.3% |
| Vending-Bench 2 | — | 3,285 |
Reasoning DeepSeek V4 Pro leads
DeepSeek LLM 67B: 16.5 (#304), DeepSeek V4 Pro: 56.5 (#24)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| Chess Puzzles | 0% | 47% |
| LMArena Hard Prompts | 1070 | 1461 |
| Epoch Capabilities Index | 110.5 | 155.31 |
| ARC-AGI-2 | — | 61.3% |
| Kagi LLM Benchmark | — | 53.5% |
| NYT Connections (extended) | — | 91.3% |
| ARC-AGI-1 | — | 90.5% |
| CritPt | — | 18% |
| Mystery Game Puzzles | — | 43% |
| DTBench | — | 93.9% |
| LMCA | — | 45.5% |
| Surface Evolver Bench | — | 40% |
| ForecastBench | — | 56.1 |
Math DeepSeek V4 Pro leads
DeepSeek LLM 67B: 8.7 (#324), DeepSeek V4 Pro: 64.8 (#30)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.8% | 98.6% |
| LMArena Math | 1108 | 1455 |
| FrontierMath (Tiers 1-3) | — | 64.6% |
| FrontierMath Tier 4 | — | 26.8% |
| MathArena Final-Answer Competitions | — | 76.6% |
| ProofBench | — | 50% |
| MATH Level 5 | 6.4% | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek LLM 67B: 7.0 (#313), DeepSeek V4 Pro: 59.5 (#31)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| GPQA Diamond | 24.6% | 91.7% |
| SimpleQA Verified | — | 52.9% |
| Vectara Hallucination Rate | — | 8.6% |
| LMArena Expert | — | 1464 |
Multilingual DeepSeek V4 Pro leads
DeepSeek LLM 67B: 29.4 (#267), DeepSeek V4 Pro: 54.4 (#45)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | 1073 | 1439 |
| LMArena Chinese | 1132 | 1486 |
| LMArena French | — | 1472 |
| LMArena German | — | 1458 |
| LMArena Japanese | — | 1445 |
| LMArena Korean | — | 1447 |
| LMArena Russian | — | 1453 |
| LMArena Spanish | — | 1458 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek LLM 67B: 55.4 (#277), DeepSeek V4 Pro: 76.1 (#47)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | 1079 | 1448 |
Long Context DeepSeek V4 Pro leads
DeepSeek LLM 67B: 33.1 (#265), DeepSeek V4 Pro: 45.0 (#51)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| LMArena Longer Query | 1092 | 1458 |
| CL-bench Life | — | 13.5% |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek LLM 67B: 31.6 (#282), DeepSeek V4 Pro: 65.5 (#46)
| Benchmark | DeepSeek LLM 67B | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | 1105 | 1451 |
| LMArena Creative Writing | 1067 | 1446 |
| LMArena Multi-Turn | 1082 | 1467 |
| EQ-Bench Creative Writing | — | 1553 |
| EQ-Bench 4 | — | 1166 |
Frequently asked questions
Is DeepSeek LLM 67B better than DeepSeek V4 Pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 24.9 on the Noometry Index.
Is DeepSeek LLM 67B or DeepSeek V4 Pro better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 31.9 in the Noometry coding category.
How many benchmarks do DeepSeek LLM 67B and DeepSeek V4 Pro share?
14 benchmarks have published results for both models. DeepSeek LLM 67B has 15 scored results on Noometry and DeepSeek V4 Pro has 48.