Model comparison
DeepSeek V4 Pro vs Llama2 70b Steerlm Chat
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 31.8 on the Noometry Index.
Last verified . 9 shared benchmarks.
DeepSeek
NVIDIASummary
- They share 9 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 7 categories and Llama2 70b Steerlm Chat in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 20.0.
Side by side
| DeepSeek V4 Pro | Llama2 70b Steerlm Chat | |
|---|---|---|
| Provider | DeepSeek | NVIDIA |
| Noometry Index | 54.3 | 31.8 |
| 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 | 9 |
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), Llama2 70b Steerlm Chat: 29.9 (#300)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Coding | 1470 | 1025 |
| 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), Llama2 70b Steerlm Chat: —
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Llama2 70b Steerlm Chat: 20.0 (#246)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Hard Prompts | 1461 | 1047 |
| 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), Llama2 70b Steerlm Chat: 31.3 (#226)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Math | 1455 | 1072 |
| 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 (#31), Llama2 70b Steerlm Chat: —
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| LMArena Expert | 1464 | — |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Llama2 70b Steerlm Chat: 28.8 (#270)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Non-English | 1439 | 1063 |
| LMArena Chinese | 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 V4 Pro: 76.1 (#47), Llama2 70b Steerlm Chat: 54.2 (#279)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Instruction Following | 1448 | 1060 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Llama2 70b Steerlm Chat: 30.4 (#288)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Longer Query | 1458 | 998 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Llama2 70b Steerlm Chat: 31.6 (#283)
| Benchmark | DeepSeek V4 Pro | Llama2 70b Steerlm Chat |
|---|---|---|
| LMArena Text | 1451 | 1098 |
| LMArena Creative Writing | 1446 | 1091 |
| LMArena Multi-Turn | 1467 | 1058 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Llama2 70b Steerlm Chat?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 31.8 on the Noometry Index.
Is DeepSeek V4 Pro or Llama2 70b Steerlm Chat better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 29.9 in the Noometry coding category.
How many benchmarks do DeepSeek V4 Pro and Llama2 70b Steerlm Chat share?
9 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Llama2 70b Steerlm Chat has 9.