Noometry

Model comparison

Command R vs Grok-2 (Dec 2024)

Grok-2 (Dec 2024) is the stronger model overall, scoring 33.7 to 31.4 on the Noometry Index.

Last verified . 25 shared benchmarks.

Command R Cohere

31.4

Rank #272 Confirmed

Grok-2 (Dec 2024) xAI

33.7

Rank #239 Confirmed

Summary

  • They share 25 benchmarks with published results for both. Command R scores higher in 2 categories and Grok-2 (Dec 2024) in 6 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Grok-2 (Dec 2024) leads 48.6 to 38.2.
  • The biggest single-benchmark swing is LiveBench Math: 19.4% for Command R and 54.9% for Grok-2 (Dec 2024).
  • Command R has downloadable open weights; the other is API-only.

Side by side

Command RGrok-2 (Dec 2024)
ProviderCoherexAI
Noometry Index31.433.7
Released2024-08-302024-08-13
WeightsOpenProprietary
Context window128K—
Max output4K—
Input $ / M tokens$0.15—
Output $ / M tokens$0.60—
Results tracked2934

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Grok-2 (Dec 2024) leads

Command R: 29.3 (#306), Grok-2 (Dec 2024): 33.3 (#258)

BenchmarkCommand RGrok-2 (Dec 2024)
LiveBench Coding17.9%46.4%
LMArena Coding11691287
WeirdML—22.2%
BigCodeBench Instruct37.1%—
BigCodeBench Complete45.2%—

Reasoning Grok-2 (Dec 2024) leads

Command R: 13.8 (#331), Grok-2 (Dec 2024): 16.9 (#299)

BenchmarkCommand RGrok-2 (Dec 2024)
LiveBench Reasoning21.9%54.8%
LMArena Hard Prompts11641272
DTBench46.4%65.2%
LiveBench Data Analysis33.3%54.5%
LiveBench27.5%54.3%
SimpleBench—22.7%
LMCA9.2%—
Epoch Capabilities Index—130.48

Math Command R leads

Command R: 28.0 (#246), Grok-2 (Dec 2024): 20.8 (#284)

BenchmarkCommand RGrok-2 (Dec 2024)
LiveBench Math19.4%54.9%
LMArena Math11551283
OTIS Mock AIME 2024-2025—11.5%
MATH Level 5—63.5%
FrontierMath (Feb 2025 set)—0.7%

Knowledge Command R leads

Command R: 31.0 (#221), Grok-2 (Dec 2024): 29.8 (#233)

BenchmarkCommand RGrok-2 (Dec 2024)
LMArena Expert11381254
GPQA Diamond—53.8%
Confabulations—20.1%
MMLU65.2%—

Multilingual Grok-2 (Dec 2024) leads

Command R: 35.7 (#245), Grok-2 (Dec 2024): 43.1 (#188)

BenchmarkCommand RGrok-2 (Dec 2024)
LMArena Non-English11741282
LMArena Chinese11821289
LMArena French11621318
LMArena German11761287
LMArena Japanese11431244
LMArena Korean11631237
LMArena Russian11741286
LMArena Spanish11511281

Instruction Following Grok-2 (Dec 2024) leads

Command R: 58.1 (#261), Grok-2 (Dec 2024): 66.9 (#202)

BenchmarkCommand RGrok-2 (Dec 2024)
LiveBench Instruction Following55.6%69.6%
LMArena Instruction Following11671270

Long Context Grok-2 (Dec 2024) leads

Command R: 36.3 (#231), Grok-2 (Dec 2024): 38.8 (#190)

BenchmarkCommand RGrok-2 (Dec 2024)
LMArena Longer Query11981276

Writing & Preference Grok-2 (Dec 2024) leads

Command R: 38.2 (#254), Grok-2 (Dec 2024): 48.6 (#198)

BenchmarkCommand RGrok-2 (Dec 2024)
LMArena Text11871305
LMArena Creative Writing11701284
LMArena Multi-Turn11631290
LiveBench Language16.7%45.6%
Short-Story Creative Writing—63.6%

Frequently asked questions

Is Command R better than Grok-2 (Dec 2024)?

Grok-2 (Dec 2024) is the stronger model overall, scoring 33.7 to 31.4 on the Noometry Index.

Is Command R or Grok-2 (Dec 2024) better for coding?

Grok-2 (Dec 2024) scores higher on coding benchmarks: 33.3 versus 29.3 in the Noometry coding category.

How many benchmarks do Command R and Grok-2 (Dec 2024) share?

25 benchmarks have published results for both models. Command R has 29 scored results on Noometry and Grok-2 (Dec 2024) has 34.

Related comparisons

Go deeper