Model comparison
Command R+ vs DeepSeek-R1
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 32.4 on the Noometry Index.
Last verified . 27 shared benchmarks.
Cohere
DeepSeekSummary
- They share 27 benchmarks with published results for both. Command R+ scores higher in 0 categories and DeepSeek-R1 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-R1 leads 61.4 to 43.5.
- The biggest single-benchmark swing is LiveBench Math: 21.3% for Command R+ and 80.7% for DeepSeek-R1.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $2.50 / $10 for Command R+.
- DeepSeek-R1 accepts more context: 164K tokens versus 128K.
- Command R+ has downloadable open weights; the other is API-only.
Side by side
| Command R+ | DeepSeek-R1 | |
|---|---|---|
| Provider | Cohere | DeepSeek |
| Noometry Index | 32.4 | 42.3 |
| Released | 2024-08-30 | 2025-01-20 |
| Weights | Open | Proprietary |
| Context window | 128K | 164K |
| Max output | 4K | 64K |
| Input $ / M tokens | $2.50 | $0.50 |
| Output $ / M tokens | $10 | $2.15 |
| Results tracked | 34 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
Command R+: 29.1 (#309), DeepSeek-R1: 46.3 (#68)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LiveBench Coding | 19.1% | 66.7% |
| LMArena Coding | 1187 | 1427 |
| Aider Polyglot | — | 71.4% |
| SciCode | — | 35.7% |
| WeirdML | — | 41.6% |
| BigCodeBench Instruct | 33.8% | — |
| BigCodeBench Complete | 41.9% | — |
| ALE-Bench | — | 804.12 |
| AlgoTune | — | 1.7 |
| HumanEval+ | 56.7% | — |
| MBPP+ | 63.5% | — |
Agentic & Tool Use Not comparable
Command R+: —, DeepSeek-R1: 30.7 (#75)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| DeepResearch Bench | — | 35.1% |
| BALROG | — | 34.9% |
| METR Time Horizons | — | 53.8% |
Reasoning DeepSeek-R1 leads
Command R+: 9.2 (#344), DeepSeek-R1: 18.6 (#278)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| SimpleBench | 17.4% | 40.8% |
| LiveBench Reasoning | 24.8% | 83.2% |
| LMArena Hard Prompts | 1186 | 1416 |
| LiveBench Data Analysis | 38.1% | 69.8% |
| Epoch Capabilities Index | 119.34 | 141.29 |
| LiveBench | 31.8% | 71.6% |
| ARC-AGI-2 | — | 1.3% |
| Kagi LLM Benchmark | — | 69.4% |
| ARC-AGI-1 | — | 21.2% |
| CritPt | — | 1.1% |
| DTBench | 54.9% | — |
| LMCA | 5% | — |
| ForecastBench | — | 60 |
Math DeepSeek-R1 leads
Command R+: 28.9 (#242), DeepSeek-R1: 43.8 (#79)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LiveBench Math | 21.3% | 80.7% |
| LMArena Math | 1188 | 1400 |
| OTIS Mock AIME 2024-2025 | — | 66.4% |
| Omni-MATH | — | 42.4% |
| MATH Level 5 | — | 96.6% |
Knowledge DeepSeek-R1 leads
Command R+: 36.4 (#169), DeepSeek-R1: 44.5 (#87)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| Vectara Hallucination Rate | 6.9% | 11.3% |
| LMArena Expert | 1174 | 1394 |
| GPQA Diamond | — | 76.3% |
| MMLU-Pro | — | 79.3% |
| Confabulations | — | 12.7% |
| GPQA (HELM) | — | 66.6% |
| MMLU | 69.4% | — |
Multilingual DeepSeek-R1 leads
Command R+: 38.6 (#227), DeepSeek-R1: 52.4 (#85)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LMArena Non-English | 1216 | 1412 |
| LMArena Chinese | 1226 | 1442 |
| LMArena French | 1209 | 1417 |
| LMArena German | 1216 | 1404 |
| LMArena Japanese | 1166 | 1391 |
| LMArena Korean | 1138 | 1360 |
| LMArena Russian | 1227 | 1423 |
| LMArena Spanish | 1189 | 1411 |
Instruction Following DeepSeek-R1 leads
Command R+: 60.0 (#254), DeepSeek-R1: 72.0 (#143)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LiveBench Instruction Following | 57.6% | 80.5% |
| LMArena Instruction Following | 1197 | 1382 |
| IFEval | — | 78.4% |
Long Context DeepSeek-R1 leads
Command R+: 37.3 (#219), DeepSeek-R1: 45.4 (#36)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LMArena Longer Query | 1230 | 1391 |
| Fiction.LiveBench | — | 75% |
Writing & Preference DeepSeek-R1 leads
Command R+: 43.5 (#228), DeepSeek-R1: 61.4 (#88)
| Benchmark | Command R+ | DeepSeek-R1 |
|---|---|---|
| LMArena Text | 1229 | 1428 |
| LMArena Creative Writing | 1235 | 1405 |
| LMArena Multi-Turn | 1213 | 1405 |
| LiveBench Language | 29.7% | 48.5% |
| Short-Story Creative Writing | — | 83% |
| EQ-Bench Creative Writing | — | 1500 |
| WildBench | — | 82.8% |
Frequently asked questions
Is Command R+ better than DeepSeek-R1?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 32.4 on the Noometry Index.
Which is cheaper, Command R+ or DeepSeek-R1?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Command R+ lists at $2.50 and $10.
Is Command R+ or DeepSeek-R1 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 29.1 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 128K.
How many benchmarks do Command R+ and DeepSeek-R1 share?
27 benchmarks have published results for both models. Command R+ has 34 scored results on Noometry and DeepSeek-R1 has 52.