Model comparison
Codellama 34b Instruct vs DeepSeek V4 Pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 30.8 on the Noometry Index.
Last verified . 10 shared benchmarks.
Meta
DeepSeekSummary
- They share 10 benchmarks with published results for both. Codellama 34b Instruct scores higher in 0 categories and DeepSeek V4 Pro in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek V4 Pro leads 65.5 to 28.2.
Side by side
| Codellama 34b Instruct | DeepSeek V4 Pro | |
|---|---|---|
| Provider | Meta | DeepSeek |
| Noometry Index | 30.8 | 54.3 |
| 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 | 14 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
Codellama 34b Instruct: 28.5 (#314), DeepSeek V4 Pro: 52.4 (#34)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Coding | 1046 | 1470 |
| SWE-bench Verified | — | 77.6% |
| FrontierCode | — | 28.6% |
| LMArena WebDev | — | 1582 |
| SciCode | — | 51% |
| WeirdML | — | 66.2% |
| BigCodeBench Instruct | 29% | — |
| BigCodeBench Complete | 37.1% | — |
| ALE-Bench | — | 1,403 |
| HumanEval+ | 43.9% | — |
| MBPP+ | 56.3% | — |
Agentic & Tool Use Not comparable
Codellama 34b Instruct: —, DeepSeek V4 Pro: 32.8 (#58)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| APEX-Agents | — | 47.3% |
| Vending-Bench 2 | — | 3,285 |
Reasoning DeepSeek V4 Pro leads
Codellama 34b Instruct: 19.6 (#255), DeepSeek V4 Pro: 56.5 (#24)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Hard Prompts | 1032 | 1461 |
| 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
Codellama 34b Instruct: 31.0 (#230), DeepSeek V4 Pro: 64.8 (#30)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Math | 1056 | 1455 |
| 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
Codellama 34b Instruct: —, DeepSeek V4 Pro: 59.5 (#31)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| GPQA Diamond | — | 91.7% |
| SimpleQA Verified | — | 52.9% |
| Vectara Hallucination Rate | — | 8.6% |
| LMArena Expert | — | 1464 |
Multilingual DeepSeek V4 Pro leads
Codellama 34b Instruct: 25.8 (#284), DeepSeek V4 Pro: 54.4 (#45)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | 1011 | 1439 |
| LMArena Chinese | 976 | 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
Codellama 34b Instruct: 52.2 (#291), DeepSeek V4 Pro: 76.1 (#47)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | 1028 | 1448 |
Long Context DeepSeek V4 Pro leads
Codellama 34b Instruct: 30.9 (#284), DeepSeek V4 Pro: 45.0 (#51)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Longer Query | 1013 | 1458 |
| CL-bench Life | — | 13.5% |
Writing & Preference DeepSeek V4 Pro leads
Codellama 34b Instruct: 28.2 (#297), DeepSeek V4 Pro: 65.5 (#46)
| Benchmark | Codellama 34b Instruct | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | 1066 | 1451 |
| LMArena Creative Writing | 1032 | 1446 |
| LMArena Multi-Turn | 1015 | 1467 |
| EQ-Bench Creative Writing | — | 1553 |
| EQ-Bench 4 | — | 1166 |
Frequently asked questions
Is Codellama 34b Instruct better than DeepSeek V4 Pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 30.8 on the Noometry Index.
Is Codellama 34b Instruct or DeepSeek V4 Pro better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 28.5 in the Noometry coding category.
How many benchmarks do Codellama 34b Instruct and DeepSeek V4 Pro share?
10 benchmarks have published results for both models. Codellama 34b Instruct has 14 scored results on Noometry and DeepSeek V4 Pro has 48.