Model comparison
Command A vs DeepSeek-V3.2-Exp
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 36.5 on the Noometry Index.
Last verified . 24 shared benchmarks.
Cohere
DeepSeekSummary
- They share 24 benchmarks with published results for both. Command A scores higher in 1 category and DeepSeek-V3.2-Exp in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V3.2-Exp leads 46.5 to 27.2.
- The biggest single-benchmark swing is Aider Polyglot: 12% for Command A and 74.2% for DeepSeek-V3.2-Exp.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $2.50 / $10 for Command A.
- Command A accepts more context: 256K tokens versus 164K.
Side by side
| Command A | DeepSeek-V3.2-Exp | |
|---|---|---|
| Provider | Cohere | DeepSeek |
| Noometry Index | 36.5 | 44.3 |
| Released | 2025-03-13 | 2025-09-29 |
| Weights | Open | Open |
| Context window | 256K | 164K |
| Max output | 8K | 66K |
| Input $ / M tokens | $2.50 | $0.26 |
| Output $ / M tokens | $10 | $0.38 |
| Results tracked | 24 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
Command A: 27.2 (#322), DeepSeek-V3.2-Exp: 46.5 (#65)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| Aider Polyglot | 12% | 74.2% |
| LMArena Coding | 1330 | 1454 |
| SWE-bench Verified (bash only) | — | 70% |
| LMArena WebDev | — | 1362 |
| SWE-bench Multilingual | — | 59% |
| SciCode | — | 38.9% |
| WeirdML | — | 39.5% |
Agentic & Tool Use Command A leads
Command A: 35.9 (#40), DeepSeek-V3.2-Exp: 32.7 (#59)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| Berkeley Function Calling Leaderboard | 57.1% | 56.7% |
| Terminal-Bench | — | 39.6% |
| APEX-Agents | — | 21.3% |
| TheAgentCompany | — | 42.9% |
| Vending-Bench 2 | — | 1,034 |
Reasoning DeepSeek-V3.2-Exp leads
Command A: 18.3 (#283), DeepSeek-V3.2-Exp: 22.1 (#208)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| Kagi LLM Benchmark | 28.8% | 52.2% |
| LMArena Hard Prompts | 1326 | 1434 |
| DTBench | 61.3% | 87.7% |
| LMCA | 10.3% | 29.1% |
| ARC-AGI-2 | — | 4% |
| NYT Connections (extended) | — | 36.7% |
| ARC-AGI-1 | — | 57% |
| CritPt | — | 2.9% |
| Chess Puzzles | — | 14% |
| Thematic Generalization | — | 65% |
| Epoch Capabilities Index | — | 146.27 |
Math DeepSeek-V3.2-Exp leads
Command A: 36.2 (#171), DeepSeek-V3.2-Exp: 41.7 (#87)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Math | 1300 | 1435 |
| MathArena Final-Answer Competitions | — | 57.7% |
| OTIS Mock AIME 2024-2025 | — | 87.8% |
| ProofBench | — | 8% |
| FrontierMath (Feb 2025 set) | — | 22.1% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge DeepSeek-V3.2-Exp leads
Command A: 37.1 (#159), DeepSeek-V3.2-Exp: 51.7 (#66)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| Vectara Hallucination Rate | 9.3% | 5.3% |
| LMArena Expert | 1295 | 1436 |
| GPQA Diamond | — | 83.4% |
Multilingual DeepSeek-V3.2-Exp leads
Command A: 45.3 (#170), DeepSeek-V3.2-Exp: 52.2 (#90)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Non-English | 1313 | 1409 |
| LMArena Chinese | 1327 | 1461 |
| LMArena French | 1351 | 1433 |
| LMArena German | 1341 | 1440 |
| LMArena Japanese | 1285 | 1374 |
| LMArena Korean | 1285 | 1371 |
| LMArena Russian | 1314 | 1424 |
| LMArena Spanish | 1347 | 1440 |
Instruction Following DeepSeek-V3.2-Exp leads
Command A: 69.1 (#177), DeepSeek-V3.2-Exp: 74.5 (#93)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Instruction Following | 1309 | 1413 |
Long Context DeepSeek-V3.2-Exp leads
Command A: 40.6 (#151), DeepSeek-V3.2-Exp: 47.6 (#16)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Longer Query | 1334 | 1428 |
| Fiction.LiveBench | — | 83.3% |
| CL-bench | — | 13.2% |
| CL-bench Life | — | 9.5% |
Writing & Preference DeepSeek-V3.2-Exp leads
Command A: 47.6 (#208), DeepSeek-V3.2-Exp: 62.4 (#77)
| Benchmark | Command A | DeepSeek-V3.2-Exp |
|---|---|---|
| LMArena Text | 1331 | 1425 |
| LMArena Creative Writing | 1319 | 1403 |
| EQ-Bench Creative Writing | 1145 | 1515 |
| LMArena Multi-Turn | 1339 | 1427 |
Frequently asked questions
Is Command A better than DeepSeek-V3.2-Exp?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 36.5 on the Noometry Index.
Which is cheaper, Command A or DeepSeek-V3.2-Exp?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Command A lists at $2.50 and $10.
Is Command A or DeepSeek-V3.2-Exp better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 27.2 in the Noometry coding category.
Which has the bigger context window?
Command A does, with 256K tokens against 164K.
How many benchmarks do Command A and DeepSeek-V3.2-Exp share?
24 benchmarks have published results for both models. Command A has 24 scored results on Noometry and DeepSeek-V3.2-Exp has 49.