Model comparison
gpt-oss-20b vs MiniMax-M2
MiniMax-M2 is the stronger model overall, scoring 37.4 to 32.5 on the Noometry Index. gpt-oss-20b costs 15× less per token, which makes it the better buy when MiniMax-M2's lead doesn't matter for your workload.
Last verified . 16 shared benchmarks.
MiniMaxSummary
- They share 16 benchmarks with published results for both. gpt-oss-20b scores higher in 1 category and MiniMax-M2 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where MiniMax-M2 leads 53.0 to 35.5.
- The biggest single-benchmark swing is Terminal-Bench: 3.4% for gpt-oss-20b and 30% for MiniMax-M2.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.
- MiniMax-M2 accepts more context: 205K tokens versus 131K.
Side by side
| gpt-oss-20b | MiniMax-M2 | |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 32.5 | 37.4 |
| Released | 2025-08-05 | 2025-10-27 |
| Weights | Open | Open |
| Context window | 131K | 205K |
| Max output | 16K | 131K |
| Input $ / M tokens | $0.018 | $0.30 |
| Output $ / M tokens | $0.09 | $1.20 |
| Results tracked | 34 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2 leads
gpt-oss-20b: 37.6 (#192), MiniMax-M2: 39.3 (#159)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Coding | 1306 | 1370 |
| SWE-bench Verified (bash only) | — | 61% |
| LMArena WebDev | — | 1297 |
| SciCode | 34.4% | — |
| WeirdML | 40.9% | — |
| ALE-Bench | 566.05 | — |
Agentic & Tool Use MiniMax-M2 leads
gpt-oss-20b: 9.3 (#154), MiniMax-M2: 25.1 (#109)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| Terminal-Bench | 3.4% | 30% |
| Vending-Bench 2 | — | 160.6 |
Reasoning Too close to call
gpt-oss-20b: 19.3 (#261), MiniMax-M2: 19.4 (#258)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 57.8% |
| LMArena Hard Prompts | 1274 | 1357 |
| NYT Connections (extended) | — | 14.8% |
| CritPt | 1.4% | — |
| Chess Puzzles | 4% | — |
| DTBench | 68% | — |
| LMCA | 14.5% | — |
| Epoch Capabilities Index | 137.82 | — |
Math gpt-oss-20b leads
gpt-oss-20b: 39.4 (#103), MiniMax-M2: 37.3 (#160)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Math | 1317 | 1352 |
| OTIS Mock AIME 2024-2025 | 65.3% | — |
| Omni-MATH | 56.5% | — |
Knowledge MiniMax-M2 leads
gpt-oss-20b: 34.6 (#195), MiniMax-M2: 37.0 (#163)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Expert | 1258 | 1337 |
| GPQA Diamond | 60.8% | — |
| MMLU-Pro | 74% | — |
| GPQA (HELM) | 59.4% | — |
Multilingual MiniMax-M2 leads
gpt-oss-20b: 42.2 (#197), MiniMax-M2: 45.3 (#171)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Non-English | 1268 | 1313 |
| LMArena Chinese | 1314 | 1366 |
| LMArena German | 1255 | 1355 |
| LMArena Russian | 1278 | 1331 |
| LMArena Spanish | 1267 | 1326 |
| LMArena French | — | 1335 |
| LMArena Japanese | 1244 | — |
| LMArena Korean | 1236 | — |
Instruction Following MiniMax-M2 leads
gpt-oss-20b: 61.8 (#240), MiniMax-M2: 70.2 (#166)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Instruction Following | 1236 | 1328 |
| IFEval | 73.2% | — |
Long Context MiniMax-M2 leads
gpt-oss-20b: 37.9 (#209), MiniMax-M2: 40.5 (#153)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Longer Query | 1250 | 1331 |
Writing & Preference MiniMax-M2 leads
gpt-oss-20b: 35.5 (#265), MiniMax-M2: 53.0 (#162)
| Benchmark | gpt-oss-20b | MiniMax-M2 |
|---|---|---|
| LMArena Text | 1287 | 1340 |
| LMArena Creative Writing | 1201 | 1286 |
| LMArena Multi-Turn | 1268 | 1361 |
| EQ-Bench Creative Writing | 666 | — |
| WildBench | 73.7% | — |
Frequently asked questions
Is gpt-oss-20b better than MiniMax-M2?
MiniMax-M2 is the stronger model overall, scoring 37.4 to 32.5 on the Noometry Index. gpt-oss-20b costs 15× less per token, which makes it the better buy when MiniMax-M2's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or MiniMax-M2?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; MiniMax-M2 lists at $0.30 and $1.20.
Is gpt-oss-20b or MiniMax-M2 better for coding?
MiniMax-M2 scores higher on coding benchmarks: 39.3 versus 37.6 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2 does, with 205K tokens against 131K.
How many benchmarks do gpt-oss-20b and MiniMax-M2 share?
16 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and MiniMax-M2 has 21.