Model comparison
DeepSeek LLM 67B vs Llama 4 Maverick
Llama 4 Maverick is the stronger model overall, scoring 30.6 to 24.8 on the Noometry Index.
Last verified . 14 shared benchmarks.
DeepSeek
MetaSummary
- They share 14 benchmarks with published results for both. DeepSeek LLM 67B scores higher in 3 categories and Llama 4 Maverick in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 4 Maverick leads 33.3 to 7.1.
- The biggest single-benchmark swing is MATH Level 5: 6.4% for DeepSeek LLM 67B and 73% for Llama 4 Maverick.
Side by side
| DeepSeek LLM 67B | Llama 4 Maverick | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 24.8 | 30.6 |
| Released | 2023-11-29 | 2025-04-05 |
| Weights | Open | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.19 |
| Output $ / M tokens | — | $0.65 |
| Results tracked | 15 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek LLM 67B leads
DeepSeek LLM 67B: 31.3 (#273), Llama 4 Maverick: 26.3 (#318)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Coding | 1096 | 1302 |
| SWE-bench Verified (bash only) | — | 21% |
| Aider Polyglot | — | 15.6% |
| SciCode | — | 33.1% |
| WeirdML | — | 24.5% |
| BigCodeBench Instruct | — | 49.7% |
| BigCodeBench Complete | — | 61.4% |
| ALE-Bench | — | 172.97 |
Agentic & Tool Use Not comparable
DeepSeek LLM 67B: —, Llama 4 Maverick: 27.9 (#86)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.3% |
Reasoning DeepSeek LLM 67B leads
DeepSeek LLM 67B: 15.9 (#300), Llama 4 Maverick: 10.1 (#337)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Hard Prompts | 1070 | 1281 |
| Epoch Capabilities Index | 110.54 | 132.2 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 27.7% |
| Kagi LLM Benchmark | — | 55.9% |
| NYT Connections (extended) | — | 8% |
| ARC-AGI-1 | — | 4.4% |
| CritPt | — | 0% |
| Chess Puzzles | 0% | — |
| EnigmaEval | — | 0.6% |
| DTBench | — | 61.9% |
| LMCA | — | 15.9% |
| ForecastBench | — | 57.5 |
Math Llama 4 Maverick leads
DeepSeek LLM 67B: 8.1 (#318), Llama 4 Maverick: 24.6 (#256)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.8% | 20.6% |
| LMArena Math | 1108 | 1299 |
| MATH Level 5 | 6.4% | 73% |
| Omni-MATH | — | 42.2% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Llama 4 Maverick leads
DeepSeek LLM 67B: 7.1 (#308), Llama 4 Maverick: 33.3 (#196)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| GPQA Diamond | 24.6% | 67% |
| Humanity's Last Exam | — | 5.7% |
| MMLU-Pro | — | 81% |
| Confabulations | — | 22.6% |
| Vectara Hallucination Rate | — | 8.2% |
| GPQA (HELM) | — | 65% |
| LMArena Expert | — | 1259 |
Multimodal Not comparable
DeepSeek LLM 67B: —, Llama 4 Maverick: 30.9 (#100)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | — | 1141 |
| GeoBench | — | 52% |
| SpatialViz-Bench | — | 31.8% |
Multilingual Llama 4 Maverick leads
DeepSeek LLM 67B: 29.5 (#262), Llama 4 Maverick: 42.4 (#190)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1073 | 1269 |
| LMArena Chinese | 1132 | 1277 |
| LMArena French | — | 1258 |
| LMArena German | — | 1291 |
| LMArena Japanese | — | 1207 |
| LMArena Korean | — | 1203 |
| LMArena Russian | — | 1286 |
| LMArena Spanish | — | 1293 |
Instruction Following Llama 4 Maverick leads
DeepSeek LLM 67B: 55.7 (#272), Llama 4 Maverick: 71.9 (#140)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Instruction Following | 1079 | 1267 |
| IFEval | — | 90.8% |
Long Context DeepSeek LLM 67B leads
DeepSeek LLM 67B: 33.1 (#260), Llama 4 Maverick: 31.4 (#274)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Longer Query | 1092 | 1280 |
| Fiction.LiveBench | — | 46.2% |
Writing & Preference Llama 4 Maverick leads
DeepSeek LLM 67B: 31.8 (#277), Llama 4 Maverick: 39.0 (#247)
| Benchmark | DeepSeek LLM 67B | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1105 | 1287 |
| LMArena Creative Writing | 1067 | 1266 |
| LMArena Multi-Turn | 1082 | 1289 |
| Short-Story Creative Writing | — | 62% |
| EQ-Bench Creative Writing | — | 860 |
| WildBench | — | 80% |
Frequently asked questions
Is DeepSeek LLM 67B better than Llama 4 Maverick?
Llama 4 Maverick is the stronger model overall, scoring 30.6 to 24.8 on the Noometry Index.
Is DeepSeek LLM 67B or Llama 4 Maverick better for coding?
DeepSeek LLM 67B scores higher on coding benchmarks: 31.3 versus 26.3 in the Noometry coding category.
How many benchmarks do DeepSeek LLM 67B and Llama 4 Maverick share?
14 benchmarks have published results for both models. DeepSeek LLM 67B has 15 scored results on Noometry and Llama 4 Maverick has 54.