Model comparison
Command A vs DeepSeek V4 Pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 36.5 on the Noometry Index.
Last verified . 22 shared benchmarks.
Cohere
DeepSeekSummary
- They share 22 benchmarks with published results for both. Command A scores higher in 1 category and DeepSeek V4 Pro in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 18.3.
- The biggest single-benchmark swing is LMCA: 10.3% for Command A and 45.5% for DeepSeek V4 Pro.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $2.50 / $10 for Command A.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 256K.
Side by side
| Command A | DeepSeek V4 Pro | |
|---|---|---|
| Provider | Cohere | DeepSeek |
| Noometry Index | 36.5 | 54.3 |
| Released | 2025-03-13 | 2026-04-24 |
| Weights | Open | Open |
| Context window | 256K | 1M |
| Max output | 8K | 393K |
| Input $ / M tokens | $2.50 | $0.66 |
| Output $ / M tokens | $10 | $1.98 |
| Results tracked | 24 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
Command A: 27.2 (#322), DeepSeek V4 Pro: 52.4 (#34)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Coding | 1330 | 1470 |
| SWE-bench Verified | — | 77.6% |
| FrontierCode | — | 28.6% |
| Aider Polyglot | 12% | — |
| LMArena WebDev | — | 1582 |
| SciCode | — | 51% |
| WeirdML | — | 66.2% |
| ALE-Bench | — | 1,403 |
Agentic & Tool Use Command A leads
Command A: 35.9 (#40), DeepSeek V4 Pro: 32.8 (#58)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| APEX-Agents | — | 47.3% |
| Berkeley Function Calling Leaderboard | 57.1% | — |
| Vending-Bench 2 | — | 3,285 |
Reasoning DeepSeek V4 Pro leads
Command A: 18.3 (#283), DeepSeek V4 Pro: 56.5 (#24)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| Kagi LLM Benchmark | 28.8% | 53.5% |
| LMArena Hard Prompts | 1326 | 1461 |
| DTBench | 61.3% | 93.9% |
| LMCA | 10.3% | 45.5% |
| ARC-AGI-2 | — | 61.3% |
| NYT Connections (extended) | — | 91.3% |
| ARC-AGI-1 | — | 90.5% |
| CritPt | — | 18% |
| Chess Puzzles | — | 47% |
| Mystery Game Puzzles | — | 43% |
| Surface Evolver Bench | — | 40% |
| Epoch Capabilities Index | — | 155.31 |
| ForecastBench | — | 56.1 |
Math DeepSeek V4 Pro leads
Command A: 36.2 (#171), DeepSeek V4 Pro: 64.8 (#30)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Math | 1300 | 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 DeepSeek V4 Pro leads
Command A: 37.1 (#159), DeepSeek V4 Pro: 59.5 (#31)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| Vectara Hallucination Rate | 9.3% | 8.6% |
| LMArena Expert | 1295 | 1464 |
| GPQA Diamond | — | 91.7% |
| SimpleQA Verified | — | 52.9% |
Multilingual DeepSeek V4 Pro leads
Command A: 45.3 (#170), DeepSeek V4 Pro: 54.4 (#45)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | 1313 | 1439 |
| LMArena Chinese | 1327 | 1486 |
| LMArena French | 1351 | 1472 |
| LMArena German | 1341 | 1458 |
| LMArena Japanese | 1285 | 1445 |
| LMArena Korean | 1285 | 1447 |
| LMArena Russian | 1314 | 1453 |
| LMArena Spanish | 1347 | 1458 |
Instruction Following DeepSeek V4 Pro leads
Command A: 69.1 (#177), DeepSeek V4 Pro: 76.1 (#47)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | 1309 | 1448 |
Long Context DeepSeek V4 Pro leads
Command A: 40.6 (#151), DeepSeek V4 Pro: 45.0 (#51)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Longer Query | 1334 | 1458 |
| CL-bench Life | — | 13.5% |
Writing & Preference DeepSeek V4 Pro leads
Command A: 47.6 (#208), DeepSeek V4 Pro: 65.5 (#46)
| Benchmark | Command A | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | 1331 | 1451 |
| LMArena Creative Writing | 1319 | 1446 |
| EQ-Bench Creative Writing | 1145 | 1553 |
| LMArena Multi-Turn | 1339 | 1467 |
| EQ-Bench 4 | — | 1166 |
Frequently asked questions
Is Command A better than DeepSeek V4 Pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 36.5 on the Noometry Index.
Which is cheaper, Command A or DeepSeek V4 Pro?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Command A lists at $2.50 and $10.
Is Command A or DeepSeek V4 Pro better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 27.2 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 256K.
How many benchmarks do Command A and DeepSeek V4 Pro share?
22 benchmarks have published results for both models. Command A has 24 scored results on Noometry and DeepSeek V4 Pro has 48.