Model comparison
DeepSeek LLM 67B vs DeepSeek-R1
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 24.9 on the Noometry Index.
Last verified . 14 shared benchmarks.
DeepSeekSummary
- They share 14 benchmarks with published results for both. DeepSeek LLM 67B scores higher in 0 categories and DeepSeek-R1 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1 leads 44.5 to 7.0.
- The biggest single-benchmark swing is MATH Level 5: 6.4% for DeepSeek LLM 67B and 96.6% for DeepSeek-R1.
- DeepSeek LLM 67B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek LLM 67B | DeepSeek-R1 | |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Noometry Index | 24.9 | 42.3 |
| Released | 2023-11-29 | 2025-01-20 |
| Weights | Open | Proprietary |
| Context window | — | 164K |
| Max output | — | 64K |
| Input $ / M tokens | — | $0.50 |
| Output $ / M tokens | — | $2.15 |
| Results tracked | 15 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek LLM 67B: 31.9 (#278), DeepSeek-R1: 46.3 (#68)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Coding | 1096 | 1427 |
| Aider Polyglot | — | 71.4% |
| SciCode | — | 35.7% |
| WeirdML | — | 41.6% |
| LiveBench Coding | — | 66.7% |
| ALE-Bench | — | 804.12 |
| AlgoTune | — | 1.7 |
Agentic & Tool Use Not comparable
DeepSeek LLM 67B: —, DeepSeek-R1: 30.7 (#75)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| DeepResearch Bench | — | 35.1% |
| BALROG | — | 34.9% |
| METR Time Horizons | — | 53.8% |
Reasoning DeepSeek-R1 leads
DeepSeek LLM 67B: 16.5 (#304), DeepSeek-R1: 18.6 (#278)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Hard Prompts | 1070 | 1416 |
| Epoch Capabilities Index | 110.5 | 141.29 |
| ARC-AGI-2 | — | 1.3% |
| SimpleBench | — | 40.8% |
| Kagi LLM Benchmark | — | 69.4% |
| ARC-AGI-1 | — | 21.2% |
| CritPt | — | 1.1% |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | — | 83.2% |
| LiveBench Data Analysis | — | 69.8% |
| ForecastBench | — | 60 |
| LiveBench | — | 71.6% |
Math DeepSeek-R1 leads
DeepSeek LLM 67B: 8.7 (#324), DeepSeek-R1: 43.8 (#79)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.8% | 66.4% |
| LMArena Math | 1108 | 1400 |
| MATH Level 5 | 6.4% | 96.6% |
| Omni-MATH | — | 42.4% |
| LiveBench Math | — | 80.7% |
Knowledge DeepSeek-R1 leads
DeepSeek LLM 67B: 7.0 (#313), DeepSeek-R1: 44.5 (#87)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| GPQA Diamond | 24.6% | 76.3% |
| MMLU-Pro | — | 79.3% |
| Confabulations | — | 12.7% |
| Vectara Hallucination Rate | — | 11.3% |
| GPQA (HELM) | — | 66.6% |
| LMArena Expert | — | 1394 |
Multilingual DeepSeek-R1 leads
DeepSeek LLM 67B: 29.4 (#267), DeepSeek-R1: 52.4 (#85)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Non-English | 1073 | 1412 |
| LMArena Chinese | 1132 | 1442 |
| LMArena French | — | 1417 |
| LMArena German | — | 1404 |
| LMArena Japanese | — | 1391 |
| LMArena Korean | — | 1360 |
| LMArena Russian | — | 1423 |
| LMArena Spanish | — | 1411 |
Instruction Following DeepSeek-R1 leads
DeepSeek LLM 67B: 55.4 (#277), DeepSeek-R1: 72.0 (#143)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Instruction Following | 1079 | 1382 |
| LiveBench Instruction Following | — | 80.5% |
| IFEval | — | 78.4% |
Long Context DeepSeek-R1 leads
DeepSeek LLM 67B: 33.1 (#265), DeepSeek-R1: 45.4 (#36)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Longer Query | 1092 | 1391 |
| Fiction.LiveBench | — | 75% |
Writing & Preference DeepSeek-R1 leads
DeepSeek LLM 67B: 31.6 (#282), DeepSeek-R1: 61.4 (#88)
| Benchmark | DeepSeek LLM 67B | DeepSeek-R1 |
|---|---|---|
| LMArena Text | 1105 | 1428 |
| LMArena Creative Writing | 1067 | 1405 |
| LMArena Multi-Turn | 1082 | 1405 |
| Short-Story Creative Writing | — | 83% |
| EQ-Bench Creative Writing | — | 1500 |
| WildBench | — | 82.8% |
| LiveBench Language | — | 48.5% |
Frequently asked questions
Is DeepSeek LLM 67B better than DeepSeek-R1?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 24.9 on the Noometry Index.
Is DeepSeek LLM 67B or DeepSeek-R1 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 31.9 in the Noometry coding category.
How many benchmarks do DeepSeek LLM 67B and DeepSeek-R1 share?
14 benchmarks have published results for both models. DeepSeek LLM 67B has 15 scored results on Noometry and DeepSeek-R1 has 52.