Model comparison
gpt-oss-20b vs Grok-3 mini
Grok-3 mini is the stronger model overall, scoring 41.0 to 32.1 on the Noometry Index.
Last verified . 26 shared benchmarks.
Summary
- They share 26 benchmarks with published results for both. gpt-oss-20b scores higher in 1 category and Grok-3 mini in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Grok-3 mini leads 52.7 to 35.8.
- The biggest single-benchmark swing is Omni-MATH: 56.5% for gpt-oss-20b and 31.8% for Grok-3 mini.
- gpt-oss-20b has downloadable open weights; the other is API-only.
Side by side
| gpt-oss-20b | Grok-3 mini | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 32.1 | 41.0 |
| Released | 2025-08-05 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | 131K | — |
| Max output | 16K | — |
| Input $ / M tokens | $0.018 | — |
| Output $ / M tokens | $0.09 | — |
| Results tracked | 34 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok-3 mini leads
gpt-oss-20b: 37.1 (#185), Grok-3 mini: 40.4 (#126)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| WeirdML | 40.9% | 42.6% |
| LMArena Coding | 1307 | 1380 |
| Aider Polyglot | — | 49.3% |
| SciCode | 34.4% | — |
| ALE-Bench | 566.05 | — |
Agentic & Tool Use Not comparable
gpt-oss-20b: 9.2 (#146), Grok-3 mini: —
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| Terminal-Bench | 3.4% | — |
Reasoning gpt-oss-20b leads
gpt-oss-20b: 18.8 (#251), Grok-3 mini: 13.8 (#323)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 61.3% |
| LMArena Hard Prompts | 1274 | 1374 |
| Epoch Capabilities Index | 137.81 | 140.35 |
| ARC-AGI-2 | — | 0.4% |
| ARC-AGI-1 | — | 16.5% |
| CritPt | 1.4% | — |
| Chess Puzzles | 4% | — |
| DTBench | 68% | — |
| LMCA | 14.5% | — |
Math Grok-3 mini leads
gpt-oss-20b: 37.6 (#93), Grok-3 mini: 40.3 (#78)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | 77.8% |
| Omni-MATH | 56.5% | 31.8% |
| LMArena Math | 1317 | 1386 |
| MATH Level 5 | — | 90.9% |
| FrontierMath (Feb 2025 set) | — | 5.9% |
Knowledge Grok-3 mini leads
gpt-oss-20b: 34.6 (#186), Grok-3 mini: 46.3 (#75)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| GPQA Diamond | 60.8% | 76.3% |
| MMLU-Pro | 74% | 79.9% |
| GPQA (HELM) | 59.4% | 67.5% |
| LMArena Expert | 1258 | 1395 |
| Confabulations | — | 10.8% |
Multilingual Grok-3 mini leads
gpt-oss-20b: 42.4 (#192), Grok-3 mini: 48.3 (#141)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| LMArena Non-English | 1269 | 1352 |
| LMArena Chinese | 1311 | 1387 |
| LMArena German | 1249 | 1350 |
| LMArena Japanese | 1245 | 1343 |
| LMArena Korean | 1239 | 1336 |
| LMArena Russian | 1279 | 1352 |
| LMArena Spanish | 1267 | 1379 |
| LMArena French | — | 1357 |
Instruction Following Grok-3 mini leads
gpt-oss-20b: 62.0 (#235), Grok-3 mini: 78.7 (#8)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| IFEval | 73.2% | 95.1% |
| LMArena Instruction Following | 1236 | 1356 |
Long Context Grok-3 mini leads
gpt-oss-20b: 37.9 (#203), Grok-3 mini: 41.0 (#142)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| LMArena Longer Query | 1251 | 1372 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference Grok-3 mini leads
gpt-oss-20b: 35.8 (#260), Grok-3 mini: 52.7 (#163)
| Benchmark | gpt-oss-20b | Grok-3 mini |
|---|---|---|
| LMArena Text | 1288 | 1370 |
| LMArena Creative Writing | 1201 | 1342 |
| WildBench | 73.7% | 65.1% |
| LMArena Multi-Turn | 1268 | 1355 |
| Short-Story Creative Writing | — | 73.5% |
| EQ-Bench Creative Writing | 666 | — |
Frequently asked questions
Is gpt-oss-20b better than Grok-3 mini?
Grok-3 mini is the stronger model overall, scoring 41.0 to 32.1 on the Noometry Index.
Is gpt-oss-20b or Grok-3 mini better for coding?
Grok-3 mini scores higher on coding benchmarks: 40.4 versus 37.1 in the Noometry coding category.
How many benchmarks do gpt-oss-20b and Grok-3 mini share?
26 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Grok-3 mini has 35.