Model comparison
Gemini 2.5 Flash vs gpt-oss-20b
Gemini 2.5 Flash is the stronger model overall, scoring 39.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 24× less per token, which makes it the better buy when Gemini 2.5 Flash's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
GoogleSummary
- They share 31 benchmarks with published results for both. Gemini 2.5 Flash scores higher in 7 categories and gpt-oss-20b in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Gemini 2.5 Flash leads 30.8 to 9.3.
- The biggest single-benchmark swing is GPQA (HELM): 39% for Gemini 2.5 Flash and 59.4% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.30 / $2.50 for Gemini 2.5 Flash.
- Gemini 2.5 Flash accepts more context: 1.05M tokens versus 131K.
- gpt-oss-20b has downloadable open weights; the other is API-only.
Side by side
| Gemini 2.5 Flash | gpt-oss-20b | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 39.3 | 32.5 |
| Released | 2025-04-17 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 131K |
| Max output | 66K | 16K |
| Input $ / M tokens | $0.30 | $0.018 |
| Output $ / M tokens | $2.50 | $0.09 |
| Results tracked | 54 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-20b leads
Gemini 2.5 Flash: 35.8 (#220), gpt-oss-20b: 37.6 (#192)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| WeirdML | 41.9% | 40.9% |
| LMArena Coding | 1424 | 1306 |
| ALE-Bench | 661.88 | 566.05 |
| SWE-bench Verified (bash only) | 28.7% | — |
| Aider Polyglot | 55.1% | — |
| SciCode | — | 34.4% |
Agentic & Tool Use Gemini 2.5 Flash leads
Gemini 2.5 Flash: 30.8 (#74), gpt-oss-20b: 9.3 (#154)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | 17.1% | 3.4% |
| Berkeley Function Calling Leaderboard | 56.2% | — |
| TheAgentCompany | 41.1% | — |
| BALROG | 33.5% | — |
| Vending-Bench 2 | 548.84 | — |
Reasoning gpt-oss-20b leads
Gemini 2.5 Flash: 18.1 (#286), gpt-oss-20b: 19.3 (#261)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 56.8% | 53.2% |
| CritPt | 1.1% | 1.4% |
| LMArena Hard Prompts | 1422 | 1274 |
| DTBench | 76.5% | 68% |
| LMCA | 27.5% | 14.5% |
| Epoch Capabilities Index | 143.03 | 137.82 |
| ARC-AGI-2 | 2.5% | — |
| SimpleBench | 41.2% | — |
| ARC-AGI-1 | 33.3% | — |
| Chess Puzzles | — | 4% |
| EnigmaEval | 2.7% | — |
| ForecastBench | 60.6 | — |
Math Too close to call
Gemini 2.5 Flash: 39.9 (#98), gpt-oss-20b: 39.4 (#103)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.1% | 65.3% |
| Omni-MATH | 38.5% | 56.5% |
| LMArena Math | 1415 | 1317 |
| FrontierMath (Feb 2025 set) | 4.8% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge Gemini 2.5 Flash leads
Gemini 2.5 Flash: 36.4 (#168), gpt-oss-20b: 34.6 (#195)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| MMLU-Pro | 63.9% | 74% |
| GPQA (HELM) | 39% | 59.4% |
| LMArena Expert | 1426 | 1258 |
| GPQA Diamond | — | 60.8% |
| Humanity's Last Exam | 12.1% | — |
| Confabulations | 16.8% | — |
| Vectara Hallucination Rate | 7.8% | — |
Multimodal Not comparable
Gemini 2.5 Flash: 41.8 (#32), gpt-oss-20b: —
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Vision | 1253 | — |
| GeoBench | 76% | — |
| VPCT | 46.2% | — |
| SpatialViz-Bench | 36.9% | — |
Multilingual Gemini 2.5 Flash leads
Gemini 2.5 Flash: 52.3 (#88), gpt-oss-20b: 42.2 (#197)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1409 | 1268 |
| LMArena Chinese | 1450 | 1314 |
| LMArena German | 1418 | 1255 |
| LMArena Japanese | 1405 | 1244 |
| LMArena Korean | 1385 | 1236 |
| LMArena Russian | 1415 | 1278 |
| LMArena Spanish | 1421 | 1267 |
| LMArena French | 1433 | — |
Instruction Following Gemini 2.5 Flash leads
Gemini 2.5 Flash: 75.7 (#54), gpt-oss-20b: 61.8 (#240)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| IFEval | 89.8% | 73.2% |
| LMArena Instruction Following | 1405 | 1236 |
Long Context Gemini 2.5 Flash leads
Gemini 2.5 Flash: 47.5 (#17), gpt-oss-20b: 37.9 (#209)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1419 | 1250 |
| Fiction.LiveBench | 77.8% | — |
Writing & Preference Gemini 2.5 Flash leads
Gemini 2.5 Flash: 53.8 (#157), gpt-oss-20b: 35.5 (#265)
| Benchmark | Gemini 2.5 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1417 | 1287 |
| LMArena Creative Writing | 1400 | 1201 |
| EQ-Bench Creative Writing | 1137 | 666 |
| WildBench | 81.7% | 73.7% |
| LMArena Multi-Turn | 1408 | 1268 |
| Short-Story Creative Writing | 76.5% | — |
Frequently asked questions
Is Gemini 2.5 Flash better than gpt-oss-20b?
Gemini 2.5 Flash is the stronger model overall, scoring 39.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 24× less per token, which makes it the better buy when Gemini 2.5 Flash's lead doesn't matter for your workload.
Which is cheaper, Gemini 2.5 Flash or gpt-oss-20b?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Gemini 2.5 Flash lists at $0.30 and $2.50.
Is Gemini 2.5 Flash or gpt-oss-20b better for coding?
gpt-oss-20b scores higher on coding benchmarks: 37.6 versus 35.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 2.5 Flash does, with 1.05M tokens against 131K.
How many benchmarks do Gemini 2.5 Flash and gpt-oss-20b share?
31 benchmarks have published results for both models. Gemini 2.5 Flash has 54 scored results on Noometry and gpt-oss-20b has 34.