Model comparison
gpt-oss-20b vs Mercury 2
Mercury 2 is the stronger model overall, scoring 39.1 to 32.5 on the Noometry Index. gpt-oss-20b costs 10× less per token, which makes it the better buy when Mercury 2's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
Summary
- They share 15 benchmarks with published results for both. gpt-oss-20b scores higher in 1 category and Mercury 2 in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mercury 2 leads 53.8 to 35.5.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.25 / $0.75 for Mercury 2.
- gpt-oss-20b accepts more context: 131K tokens versus 128K.
- gpt-oss-20b has downloadable open weights; the other is API-only.
Side by side
| gpt-oss-20b | Mercury 2 | |
|---|---|---|
| Provider | OpenAI | Inception |
| Noometry Index | 32.5 | 39.1 |
| Released | 2025-08-05 | 2026-02-20 |
| Weights | Open | Proprietary |
| Context window | 131K | 128K |
| Max output | 16K | 50K |
| Input $ / M tokens | $0.018 | $0.25 |
| Output $ / M tokens | $0.09 | $0.75 |
| Results tracked | 34 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-20b leads
gpt-oss-20b: 37.6 (#192), Mercury 2: 33.5 (#255)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| SciCode | 34.4% | 38.7% |
| WeirdML | 40.9% | 43.2% |
| LMArena Coding | 1306 | 1391 |
| ALE-Bench | 566.05 | 785.58 |
| LMArena WebDev | — | 1171 |
Agentic & Tool Use Not comparable
gpt-oss-20b: 9.3 (#154), Mercury 2: —
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| Terminal-Bench | 3.4% | — |
Reasoning Mercury 2 leads
gpt-oss-20b: 19.3 (#261), Mercury 2: 23.8 (#170)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| CritPt | 1.4% | 0.8% |
| LMArena Hard Prompts | 1274 | 1362 |
| Kagi LLM Benchmark | 53.2% | — |
| Chess Puzzles | 4% | — |
| DTBench | 68% | — |
| LMCA | 14.5% | — |
| Epoch Capabilities Index | 137.82 | — |
Math Not comparable
gpt-oss-20b: 39.4 (#103), Mercury 2: —
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | — |
| Omni-MATH | 56.5% | — |
| LMArena Math | 1317 | — |
Knowledge Mercury 2 leads
gpt-oss-20b: 34.6 (#195), Mercury 2: 36.2 (#172)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| LMArena Expert | 1258 | 1358 |
| GPQA Diamond | 60.8% | — |
| MMLU-Pro | 74% | — |
| Vectara Hallucination Rate | — | 12.3% |
| GPQA (HELM) | 59.4% | — |
Multilingual Mercury 2 leads
gpt-oss-20b: 42.2 (#197), Mercury 2: 46.6 (#157)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| LMArena Non-English | 1268 | 1331 |
| LMArena Chinese | 1314 | 1417 |
| LMArena Russian | 1278 | 1304 |
| LMArena German | 1255 | — |
| LMArena Japanese | 1244 | — |
| LMArena Korean | 1236 | — |
| LMArena Spanish | 1267 | — |
Instruction Following Mercury 2 leads
gpt-oss-20b: 61.8 (#240), Mercury 2: 70.2 (#165)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| LMArena Instruction Following | 1236 | 1329 |
| IFEval | 73.2% | — |
Long Context Mercury 2 leads
gpt-oss-20b: 37.9 (#209), Mercury 2: 40.5 (#154)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| LMArena Longer Query | 1250 | 1330 |
Writing & Preference Mercury 2 leads
gpt-oss-20b: 35.5 (#265), Mercury 2: 53.8 (#155)
| Benchmark | gpt-oss-20b | Mercury 2 |
|---|---|---|
| LMArena Text | 1287 | 1355 |
| LMArena Creative Writing | 1201 | 1289 |
| LMArena Multi-Turn | 1268 | 1358 |
| EQ-Bench Creative Writing | 666 | — |
| WildBench | 73.7% | — |
Frequently asked questions
Is gpt-oss-20b better than Mercury 2?
Mercury 2 is the stronger model overall, scoring 39.1 to 32.5 on the Noometry Index. gpt-oss-20b costs 10× less per token, which makes it the better buy when Mercury 2's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or Mercury 2?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Mercury 2 lists at $0.25 and $0.75.
Is gpt-oss-20b or Mercury 2 better for coding?
gpt-oss-20b scores higher on coding benchmarks: 37.6 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
gpt-oss-20b does, with 131K tokens against 128K.
How many benchmarks do gpt-oss-20b and Mercury 2 share?
15 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Mercury 2 has 17.