Model comparison
C4ai Aya Expanse 32b vs o4-mini
o4-mini is the stronger model overall, scoring 41.6 to 35.3 on the Noometry Index.
Last verified . 18 shared benchmarks.
CohereSummary
- They share 18 benchmarks with published results for both. C4ai Aya Expanse 32b scores higher in 0 categories and o4-mini in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where o4-mini leads 75.4 to 62.9.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 10.9% for C4ai Aya Expanse 32b and 18.6% for o4-mini.
- o4-mini accepts more context: 200K tokens versus 128K.
- C4ai Aya Expanse 32b has downloadable open weights; the other is API-only.
Side by side
| C4ai Aya Expanse 32b | o4-mini | |
|---|---|---|
| Provider | Cohere | OpenAI |
| Noometry Index | 35.3 | 41.6 |
| Released | 2024-10-24 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 128K | 200K |
| Max output | 4K | 100K |
| Input $ / M tokens | — | $1.10 |
| Output $ / M tokens | — | $4.40 |
| Results tracked | 18 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
C4ai Aya Expanse 32b: 34.3 (#227), o4-mini: 41.6 (#110)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Coding | 1197 | 1369 |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| GSO | — | 3.6% |
| WeirdML | — | 52.6% |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |
Agentic & Tool Use Not comparable
C4ai Aya Expanse 32b: —, o4-mini: 32.3 (#57)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
| METR Time Horizons | — | 63.9% |
Reasoning o4-mini leads
C4ai Aya Expanse 32b: 22.3 (#183), o4-mini: 24.6 (#143)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Hard Prompts | 1193 | 1351 |
| ARC-AGI-2 | — | 6.1% |
| SimpleBench | — | 38.7% |
| Kagi LLM Benchmark | — | 67.6% |
| ARC-AGI-1 | — | 58.7% |
| CritPt | — | 0.6% |
| Chess Puzzles | — | 26% |
| EnigmaEval | — | 9.2% |
| Mystery Game Puzzles | — | 5% |
| DTBench | — | 77.6% |
| LMCA | — | 26.5% |
| Epoch Capabilities Index | — | 145.63 |
| ForecastBench | — | 61.8 |
Math o4-mini leads
C4ai Aya Expanse 32b: 31.4 (#195), o4-mini: 40.2 (#79)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Math | 1200 | 1390 |
| FrontierMath (Tiers 1-3) | — | 36.1% |
| FrontierMath Tier 4 | — | 4.9% |
| OTIS Mock AIME 2024-2025 | — | 81.7% |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge o4-mini leads
C4ai Aya Expanse 32b: 32.6 (#202), o4-mini: 43.5 (#83)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| Vectara Hallucination Rate | 10.9% | 18.6% |
| LMArena Expert | 1182 | 1344 |
| GPQA Diamond | — | 79.6% |
| Humanity's Last Exam | — | 18.1% |
| SimpleQA Verified | — | 19.6% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| GPQA (HELM) | — | 73.5% |
Multimodal Not comparable
C4ai Aya Expanse 32b: —, o4-mini: 39.5 (#46)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Vision | — | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
Multilingual o4-mini leads
C4ai Aya Expanse 32b: 38.5 (#225), o4-mini: 47.2 (#149)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Non-English | 1213 | 1336 |
| LMArena Chinese | 1210 | 1355 |
| LMArena French | 1247 | 1362 |
| LMArena German | 1199 | 1336 |
| LMArena Japanese | 1163 | 1308 |
| LMArena Korean | 1158 | 1313 |
| LMArena Russian | 1227 | 1334 |
| LMArena Spanish | 1193 | 1346 |
Instruction Following o4-mini leads
C4ai Aya Expanse 32b: 62.9 (#232), o4-mini: 75.4 (#63)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1196 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
C4ai Aya Expanse 32b: 37.2 (#215), o4-mini: 45.5 (#31)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Longer Query | 1228 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference o4-mini leads
C4ai Aya Expanse 32b: 42.4 (#230), o4-mini: 54.2 (#148)
| Benchmark | C4ai Aya Expanse 32b | o4-mini |
|---|---|---|
| LMArena Text | 1224 | 1353 |
| LMArena Creative Writing | 1200 | 1294 |
| LMArena Multi-Turn | 1190 | 1350 |
| Short-Story Creative Writing | — | 75% |
| WildBench | — | 85.4% |
Frequently asked questions
Is C4ai Aya Expanse 32b better than o4-mini?
o4-mini is the stronger model overall, scoring 41.6 to 35.3 on the Noometry Index.
Is C4ai Aya Expanse 32b or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 41.6 versus 34.3 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 128K.
How many benchmarks do C4ai Aya Expanse 32b and o4-mini share?
18 benchmarks have published results for both models. C4ai Aya Expanse 32b has 18 scored results on Noometry and o4-mini has 60.