Model comparison
Command A vs o3
o3 is the stronger model overall, scoring 47.5 to 36.5 on the Noometry Index.
Last verified . 23 shared benchmarks.
CohereSummary
- They share 23 benchmarks with published results for both. Command A scores higher in 1 category and o3 in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where o3 leads 46.8 to 27.2.
- The biggest single-benchmark swing is Aider Polyglot: 12% for Command A and 81.3% for o3.
- o3 is cheaper at $2 / $8 per million input/output tokens, against $2.50 / $10 for Command A.
- Command A accepts more context: 256K tokens versus 200K.
- Command A has downloadable open weights; the other is API-only.
Side by side
| Command A | o3 | |
|---|---|---|
| Provider | Cohere | OpenAI |
| Noometry Index | 36.5 | 47.5 |
| Released | 2025-03-13 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 256K | 200K |
| Max output | 8K | 100K |
| Input $ / M tokens | $2.50 | $2 |
| Output $ / M tokens | $10 | $8 |
| Results tracked | 24 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Command A: 27.2 (#322), o3: 46.8 (#64)
| Benchmark | Command A | o3 |
|---|---|---|
| Aider Polyglot | 12% | 81.3% |
| LMArena Coding | 1330 | 1408 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| GSO | — | 8.8% |
| WeirdML | — | 52.4% |
| CadEval | — | 74% |
| ALE-Bench | — | 933.55 |
Agentic & Tool Use Command A leads
Command A: 35.9 (#40), o3: 34.5 (#44)
| Benchmark | Command A | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 57.1% | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
Reasoning o3 leads
Command A: 18.3 (#283), o3: 32.0 (#78)
| Benchmark | Command A | o3 |
|---|---|---|
| Kagi LLM Benchmark | 28.8% | 67.6% |
| LMArena Hard Prompts | 1326 | 1402 |
| DTBench | 61.3% | 84.8% |
| LMCA | 10.3% | 39.7% |
| ARC-AGI-2 | — | 6.5% |
| SimpleBench | — | 53.1% |
| ARC-AGI-1 | — | 60.8% |
| CritPt | — | 1.4% |
| Chess Puzzles | — | 38% |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
| Epoch Capabilities Index | — | 146.86 |
| ForecastBench | — | 62.5 |
Math o3 leads
Command A: 36.2 (#171), o3: 50.2 (#58)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Math | 1300 | 1426 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| OTIS Mock AIME 2024-2025 | — | 84.4% |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge o3 leads
Command A: 37.1 (#159), o3: 54.6 (#52)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Expert | 1295 | 1402 |
| GPQA Diamond | — | 81.8% |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| Vectara Hallucination Rate | 9.3% | — |
| GPQA (HELM) | — | 75.3% |
Multimodal Not comparable
Command A: —, o3: 41.4 (#36)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual o3 leads
Command A: 45.3 (#170), o3: 51.7 (#105)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Non-English | 1313 | 1401 |
| LMArena Chinese | 1327 | 1437 |
| LMArena French | 1351 | 1430 |
| LMArena German | 1341 | 1420 |
| LMArena Japanese | 1285 | 1403 |
| LMArena Korean | 1285 | 1370 |
| LMArena Russian | 1314 | 1406 |
| LMArena Spanish | 1347 | 1395 |
Instruction Following o3 leads
Command A: 69.1 (#177), o3: 72.8 (#127)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Instruction Following | 1309 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
Command A: 40.6 (#151), o3: 53.3 (#6)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Longer Query | 1334 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Command A: 47.6 (#208), o3: 63.5 (#64)
| Benchmark | Command A | o3 |
|---|---|---|
| LMArena Text | 1331 | 1410 |
| LMArena Creative Writing | 1319 | 1359 |
| EQ-Bench Creative Writing | 1145 | 1676 |
| LMArena Multi-Turn | 1339 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
Frequently asked questions
Is Command A better than o3?
o3 is the stronger model overall, scoring 47.5 to 36.5 on the Noometry Index.
Which is cheaper, Command A or o3?
o3 is cheaper. It lists at $2 per million input tokens and $8 per million output tokens; Command A lists at $2.50 and $10.
Is Command A or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 27.2 in the Noometry coding category.
Which has the bigger context window?
Command A does, with 256K tokens against 200K.
How many benchmarks do Command A and o3 share?
23 benchmarks have published results for both models. Command A has 24 scored results on Noometry and o3 has 63.