Model comparison
Codellama 34b Instruct vs DeepSeek-V3
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 30.8 on the Noometry Index.
Last verified . 14 shared benchmarks.
Meta
DeepSeekSummary
- They share 14 benchmarks with published results for both. Codellama 34b Instruct scores higher in 0 categories and DeepSeek-V3 in 7 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-V3 leads 57.4 to 28.2.
- The biggest single-benchmark swing is BigCodeBench Complete: 37.1% for Codellama 34b Instruct and 62.2% for DeepSeek-V3.
Side by side
| Codellama 34b Instruct | DeepSeek-V3 | |
|---|---|---|
| Provider | Meta | DeepSeek |
| Noometry Index | 30.8 | 39.5 |
| Released | — | 2024-12-26 |
| Weights | Open | Open |
| Context window | — | 164K |
| Max output | — | 164K |
| Input $ / M tokens | — | $0.24 |
| Output $ / M tokens | — | $0.90 |
| Results tracked | 14 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3 leads
Codellama 34b Instruct: 28.5 (#314), DeepSeek-V3: 42.3 (#106)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| BigCodeBench Instruct | 29% | 50% |
| LMArena Coding | 1046 | 1368 |
| BigCodeBench Complete | 37.1% | 62.2% |
| HumanEval+ | 43.9% | 86.6% |
| MBPP+ | 56.3% | 73% |
| Aider Polyglot | — | 55.1% |
| SciCode | — | 35.8% |
| WeirdML | — | 36.1% |
| LiveBench Coding | — | 70.9% |
Agentic & Tool Use Not comparable
Codellama 34b Instruct: —, DeepSeek-V3: —
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| METR Time Horizons | — | 49.6% |
Reasoning Too close to call
Codellama 34b Instruct: 19.6 (#255), DeepSeek-V3: 20.5 (#236)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Hard Prompts | 1032 | 1365 |
| SimpleBench | — | 27.2% |
| Kagi LLM Benchmark | — | 52.3% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 65.8% |
| DTBench | — | 64.8% |
| LiveBench Data Analysis | — | 60.9% |
| LMCA | — | 15.5% |
| BIG-Bench Hard | — | 87.5% |
| Epoch Capabilities Index | — | 135.94 |
| ForecastBench | — | 59.1 |
| HellaSwag | — | 88.9% |
| LiveBench | — | 66.9% |
| PIQA | — | 84.7% |
| WinoGrande | — | 85.2% |
Math DeepSeek-V3 leads
Codellama 34b Instruct: 31.0 (#230), DeepSeek-V3: 32.1 (#219)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Math | 1056 | 1373 |
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| Omni-MATH | — | 40.3% |
| LiveBench Math | — | 73.5% |
| MATH Level 5 | — | 75.5% |
| FrontierMath (Feb 2025 set) | — | 1.7% |
Knowledge Not comparable
Codellama 34b Instruct: —, DeepSeek-V3: 37.5 (#155)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| GPQA Diamond | — | 67.6% |
| MMLU-Pro | — | 72.3% |
| Confabulations | — | 26.1% |
| Vectara Hallucination Rate | — | 6.1% |
| GPQA (HELM) | — | 53.8% |
| LMArena Expert | — | 1351 |
| ARC (AI2) Challenge | — | 95.3% |
| MMLU | — | 87.2% |
| TriviaQA | — | 82.9% |
Multilingual DeepSeek-V3 leads
Codellama 34b Instruct: 25.8 (#284), DeepSeek-V3: 48.5 (#143)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Non-English | 1011 | 1358 |
| LMArena Chinese | 976 | 1391 |
| LMArena French | — | 1385 |
| LMArena German | — | 1374 |
| LMArena Japanese | — | 1333 |
| LMArena Korean | — | 1319 |
| LMArena Russian | — | 1373 |
| LMArena Spanish | — | 1358 |
Instruction Following DeepSeek-V3 leads
Codellama 34b Instruct: 52.2 (#291), DeepSeek-V3: 72.8 (#130)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Instruction Following | 1028 | 1345 |
| LiveBench Instruction Following | — | 81.5% |
| IFEval | — | 83.2% |
Long Context DeepSeek-V3 leads
Codellama 34b Instruct: 30.9 (#284), DeepSeek-V3: 34.0 (#253)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Longer Query | 1013 | 1352 |
| Fiction.LiveBench | — | 50% |
Writing & Preference DeepSeek-V3 leads
Codellama 34b Instruct: 28.2 (#297), DeepSeek-V3: 57.4 (#130)
| Benchmark | Codellama 34b Instruct | DeepSeek-V3 |
|---|---|---|
| LMArena Text | 1066 | 1375 |
| LMArena Creative Writing | 1032 | 1364 |
| LMArena Multi-Turn | 1015 | 1389 |
| Short-Story Creative Writing | — | 77% |
| EQ-Bench Creative Writing | — | 1472 |
| WildBench | — | 83% |
| LiveBench Language | — | 49.1% |
Frequently asked questions
Is Codellama 34b Instruct better than DeepSeek-V3?
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 30.8 on the Noometry Index.
Is Codellama 34b Instruct or DeepSeek-V3 better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 28.5 in the Noometry coding category.
How many benchmarks do Codellama 34b Instruct and DeepSeek-V3 share?
14 benchmarks have published results for both models. Codellama 34b Instruct has 14 scored results on Noometry and DeepSeek-V3 has 60.