Model comparison
Codellama 34b Instruct vs DeepSeek-R1-Distill-Llama-70B
DeepSeek-R1-Distill-Llama-70B is the stronger model overall, scoring 37.4 to 30.2 on the Noometry Index.
Last verified . 2 shared benchmarks.
Meta
DeepSeekSummary
- They share 2 benchmarks with published results for both. Codellama 34b Instruct scores higher in 0 categories and DeepSeek-R1-Distill-Llama-70B in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-R1-Distill-Llama-70B leads 49.0 to 28.4.
- The biggest single-benchmark swing is BigCodeBench Complete: 37.1% for Codellama 34b Instruct and 49.9% for DeepSeek-R1-Distill-Llama-70B.
Side by side
| Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B | |
|---|---|---|
| Provider | Meta | DeepSeek |
| Noometry Index | 30.2 | 37.4 |
| Released | — | 2025-01-20 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 14 | 13 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1-Distill-Llama-70B leads
Codellama 34b Instruct: 28.0 (#310), DeepSeek-R1-Distill-Llama-70B: 36.3 (#197)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| BigCodeBench Instruct | 29% | 35.3% |
| BigCodeBench Complete | 37.1% | 49.9% |
| LiveBench Coding | — | 51.6% |
| LMArena Coding | 1046 | — |
| HumanEval+ | 43.9% | — |
| MBPP+ | 56.3% | — |
Reasoning DeepSeek-R1-Distill-Llama-70B leads
Codellama 34b Instruct: 18.8 (#250), DeepSeek-R1-Distill-Llama-70B: 24.1 (#150)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| Kagi LLM Benchmark | — | 52.3% |
| LiveBench Reasoning | — | 67.6% |
| LMArena Hard Prompts | 1032 | — |
| LiveBench Data Analysis | — | 55.9% |
| LiveBench | — | 54.5% |
Math DeepSeek-R1-Distill-Llama-70B leads
Codellama 34b Instruct: 28.5 (#227), DeepSeek-R1-Distill-Llama-70B: 34.3 (#161)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 51.4% |
| LiveBench Math | — | 58.1% |
| LMArena Math | 1056 | — |
| MATH Level 5 | — | 89.9% |
Knowledge Not comparable
Codellama 34b Instruct: —, DeepSeek-R1-Distill-Llama-70B: 30.7 (#216)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| GPQA Diamond | — | 55.7% |
Multilingual Not comparable
Codellama 34b Instruct: 25.9 (#279), DeepSeek-R1-Distill-Llama-70B: —
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| LMArena Non-English | 1011 | — |
| LMArena Chinese | 976 | — |
Instruction Following DeepSeek-R1-Distill-Llama-70B leads
Codellama 34b Instruct: 52.4 (#286), DeepSeek-R1-Distill-Llama-70B: 68.3 (#185)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| LiveBench Instruction Following | — | 69.9% |
| LMArena Instruction Following | 1028 | — |
Long Context Not comparable
Codellama 34b Instruct: 30.8 (#279), DeepSeek-R1-Distill-Llama-70B: —
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| LMArena Longer Query | 1014 | — |
Writing & Preference DeepSeek-R1-Distill-Llama-70B leads
Codellama 34b Instruct: 28.4 (#292), DeepSeek-R1-Distill-Llama-70B: 49.0 (#189)
| Benchmark | Codellama 34b Instruct | DeepSeek-R1-Distill-Llama-70B |
|---|---|---|
| LMArena Text | 1066 | — |
| LMArena Creative Writing | 1032 | — |
| LMArena Multi-Turn | 1015 | — |
| LiveBench Language | — | 23.8% |
Frequently asked questions
Is Codellama 34b Instruct better than DeepSeek-R1-Distill-Llama-70B?
DeepSeek-R1-Distill-Llama-70B is the stronger model overall, scoring 37.4 to 30.2 on the Noometry Index.
Is Codellama 34b Instruct or DeepSeek-R1-Distill-Llama-70B better for coding?
DeepSeek-R1-Distill-Llama-70B scores higher on coding benchmarks: 36.3 versus 28.0 in the Noometry coding category.
How many benchmarks do Codellama 34b Instruct and DeepSeek-R1-Distill-Llama-70B share?
2 benchmarks have published results for both models. Codellama 34b Instruct has 14 scored results on Noometry and DeepSeek-R1-Distill-Llama-70B has 13.