Model comparison
Command R+ vs Llama 3.1-70B
Command R+ is the stronger model overall, scoring 32.4 to 29.6 on the Noometry Index. Llama 3.1-70B costs 11× less per token, which makes it the better buy when Command R+'s lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
Cohere
MetaSummary
- They share 23 benchmarks with published results for both. Command R+ scores higher in 3 categories and Llama 3.1-70B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Command R+ leads 28.9 to 13.5.
- The biggest single-benchmark swing is BigCodeBench Complete: 41.9% for Command R+ and 54.8% for Llama 3.1-70B.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $2.50 / $10 for Command R+.
Side by side
| Command R+ | Llama 3.1-70B | |
|---|---|---|
| Provider | Cohere | Meta |
| Noometry Index | 32.4 | 29.6 |
| Released | 2024-08-30 | 2024-07-23 |
| Weights | Open | Open |
| Context window | 128K | 128K |
| Max output | 4K | 4K |
| Input $ / M tokens | $2.50 | $0.40 |
| Output $ / M tokens | $10 | $0.40 |
| Results tracked | 34 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3.1-70B leads
Command R+: 29.1 (#309), Llama 3.1-70B: 30.3 (#296)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| BigCodeBench Instruct | 33.8% | 46.1% |
| LMArena Coding | 1187 | 1260 |
| BigCodeBench Complete | 41.9% | 54.8% |
| WeirdML | — | 9% |
| LiveBench Coding | 19.1% | — |
| HumanEval+ | 56.7% | — |
| MBPP+ | 63.5% | — |
Agentic & Tool Use Not comparable
Command R+: —, Llama 3.1-70B: 25.1 (#112)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| TheAgentCompany | — | 6.9% |
| BALROG | — | 27.9% |
Reasoning Llama 3.1-70B leads
Command R+: 9.2 (#344), Llama 3.1-70B: 21.6 (#220)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Hard Prompts | 1186 | 1241 |
| DTBench | 54.9% | 60% |
| LMCA | 5% | 14.8% |
| Epoch Capabilities Index | 119.34 | 125.92 |
| SimpleBench | 17.4% | — |
| LiveBench Reasoning | 24.8% | — |
| LiveBench Data Analysis | 38.1% | — |
| LiveBench | 31.8% | — |
Math Command R+ leads
Command R+: 28.9 (#242), Llama 3.1-70B: 13.5 (#304)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Math | 1188 | 1252 |
| OTIS Mock AIME 2024-2025 | — | 3.6% |
| Omni-MATH | — | 21% |
| LiveBench Math | 21.3% | — |
| MATH Level 5 | — | 36.7% |
Knowledge Command R+ leads
Command R+: 36.4 (#169), Llama 3.1-70B: 24.2 (#269)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Expert | 1174 | 1209 |
| MMLU | 69.4% | 80.1% |
| GPQA Diamond | — | 44.2% |
| MMLU-Pro | — | 65.3% |
| Vectara Hallucination Rate | 6.9% | — |
| GPQA (HELM) | — | 42.6% |
Multilingual Too close to call
Command R+: 38.6 (#227), Llama 3.1-70B: 38.8 (#225)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Non-English | 1216 | 1219 |
| LMArena Chinese | 1226 | 1215 |
| LMArena French | 1209 | 1261 |
| LMArena German | 1216 | 1222 |
| LMArena Japanese | 1166 | 1132 |
| LMArena Korean | 1138 | 1140 |
| LMArena Russian | 1227 | 1234 |
| LMArena Spanish | 1189 | 1253 |
Instruction Following Llama 3.1-70B leads
Command R+: 60.0 (#254), Llama 3.1-70B: 65.3 (#223)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Instruction Following | 1197 | 1231 |
| LiveBench Instruction Following | 57.6% | — |
| IFEval | — | 82.1% |
Long Context Too close to call
Command R+: 37.3 (#219), Llama 3.1-70B: 37.6 (#214)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Longer Query | 1230 | 1241 |
Writing & Preference Command R+ leads
Command R+: 43.5 (#228), Llama 3.1-70B: 35.4 (#267)
| Benchmark | Command R+ | Llama 3.1-70B |
|---|---|---|
| LMArena Text | 1229 | 1261 |
| LMArena Creative Writing | 1235 | 1232 |
| LMArena Multi-Turn | 1213 | 1256 |
| EQ-Bench Creative Writing | — | 784 |
| WildBench | — | 75.8% |
| LiveBench Language | 29.7% | — |
Frequently asked questions
Is Command R+ better than Llama 3.1-70B?
Command R+ is the stronger model overall, scoring 32.4 to 29.6 on the Noometry Index. Llama 3.1-70B costs 11× less per token, which makes it the better buy when Command R+'s lead doesn't matter for your workload.
Which is cheaper, Command R+ or Llama 3.1-70B?
Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Command R+ lists at $2.50 and $10.
Is Command R+ or Llama 3.1-70B better for coding?
Llama 3.1-70B scores higher on coding benchmarks: 30.3 versus 29.1 in the Noometry coding category.
Which has the bigger context window?
Both accept 128K tokens.
How many benchmarks do Command R+ and Llama 3.1-70B share?
23 benchmarks have published results for both models. Command R+ has 34 scored results on Noometry and Llama 3.1-70B has 35.