GPT-4.1 vs gpt-oss-20b
GPT-4.1 leads gpt-oss-20b on the Artificial Analysis Intelligence Index, 12.7 to 10.0, and gpt-oss-20b costs 53.3x less per output token ($0.15 vs $8.00 per million). Updated October 22, 2026; 60 benchmarks have a result for both.
GPT-4.1 or gpt-oss-20b: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick GPT-4.1 for a larger context window (1.05M against 131K tokens) and work measured by LiveCodeBench (Vals AI) or LiveCodeBench (Artificial Analysis run), where it leads.
- Shared results
- 60
- GPT-4.1 only
- 32
- gpt-oss-20b only
- 2
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
GPT-4.1
GPT-4.1 leads, 12.7 against 10.0.
Same board, both measured
Reasoning
GPQA Diamond
gpt-oss-20b
gpt-oss-20b leads, 68.9% against 65.4%.
Same board, both measured
Math
AIME
gpt-oss-20b
gpt-oss-20b leads, 86.0% against 39.6%.
Same board, both measured
Knowledge
MMLU Pro
GPT-4.1
GPT-4.1 leads, 80.5% against 71.6%.
Same board, both measured
Web design
Design Arena: websites
GPT-4.1
GPT-4.1 leads, 1038 against 852.
Same board, both measured
Long documents
Listed context window
GPT-4.1
GPT-4.1 takes 1.05M tokens against 131K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
gpt-oss-20b
GPT-4.1 $0.0060, gpt-oss-20b $0.0001 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only GPT-4.1 has the number, $0.37.
Directional only
At a glance
- Coding index
- GPT-4.1 not benchmarked, gpt-oss-20b not benchmarked.
- Widest category
- Math, 26.1 points apart on average over 8 shared benchmarks.
- Widest single gap
- AIME 2025 (Artificial Analysis run): GPT-4.1 34.7% against gpt-oss-20b 89.3%, 54.7 points apart on a 0 to 100 scale.
- Output price
- GPT-4.1 $8.00 against gpt-oss-20b $0.15 per million output tokens, 53.3x apart.
- Context window
- GPT-4.1 takes 1.05M tokens, gpt-oss-20b 131K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | GPT-4.1 | Bars | gpt-oss-20b | Ties |
|---|---|---|---|---|
| Coding | 3 | 4 | ||
| Intelligence | 1 | 1 | ||
| Agentic work | 0 | 2 | ||
| Reasoning | 4 | 4 | ||
| Math | 2 | 6 | ||
| Knowledge | 11 | 0 | ||
| Web design | 2 | 0 | ||
| General | 19 | 1 | ||
| All shared boards | 42 | 18 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index34
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| Aider polyglot | 52.4% | Not measured | Only GPT-4.1 measured |
| SWE-bench Verified | 39.6% | Not measured | Only GPT-4.1 measured |
| SWE-bench Multimodal | 31.1% | Not measured | Only GPT-4.1 measured |
| LiveCodeBench (Vals AI) | 54.7% | 80.4% | gpt-oss-20b |
| LiveCodeBench (Artificial Analysis run) | 45.7% | 77.7% | gpt-oss-20b |
| SciCode | Not measured | 38.9% | Only gpt-oss-20b measured |
| Terminal-Bench Hard (Artificial Analysis run) | 13.6% | 10.6% | GPT-4.1 |
| Terminal-Bench | Not measured | 3.4% | Only gpt-oss-20b measured |
| WeirdML | 39.0% | 40.9% | gpt-oss-20b |
| ALE-Bench | 558.1 | 566.0 | gpt-oss-20b |
| LMArena Text: coding | 1390 | 1307 | GPT-4.1 |
| LMArena Text: software and IT | 1391 | 1311 | GPT-4.1 |
| SciCode | 4.6% | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 26 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| AA Intelligence Index | 12.7 | 10.0 | GPT-4.1 |
| Epoch Capabilities Index | 136.7 | 137.7 | gpt-oss-20b |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 47.1% | 60.2% | gpt-oss-20b |
| Berkeley Function Calling Leaderboard (BFCL) | 54.0% | Not measured | Only GPT-4.1 measured |
| BrowseComp-Plus | 36.4% | 50.6% | gpt-oss-20b |
| Neither model has a result on the other 38 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index44
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| GPQA Diamond | 65.4% | 68.9% | gpt-oss-20b |
| MMMU Pro (Vals AI) | 72.4% | Not measured | Only GPT-4.1 measured |
| GPQA Diamond (Artificial Analysis run) | 66.6% | 68.8% | gpt-oss-20b |
| Humanity's Last Exam (Artificial Analysis run) | 4.2% | 11.0% | gpt-oss-20b |
| GPQA Diamond (Epoch AI run) | 66.9% | 60.8% | GPT-4.1 |
| ARC-AGI | 5.5% | Not measured | Only GPT-4.1 measured |
| ARC-AGI-2 | 0.4% | Not measured | Only GPT-4.1 measured |
| SimpleBench | 27.0% | Not measured | Only GPT-4.1 measured |
| Chess puzzles | 3.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: hard prompts | 1384 | 1274 | GPT-4.1 |
| LMArena Text: expert | 1364 | 1258 | GPT-4.1 |
| LMArena Text: hard prompts in English | 1393 | 1292 | GPT-4.1 |
| LMArena Vision: diagrams | 1215 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: homework | 1248 | Not measured | Only GPT-4.1 measured |
| Kagi LLM Benchmark | 52.3% | 53.2% | gpt-oss-20b |
| Neither model has a result on the other 7 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index26
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| AIME | 39.6% | 86.0% | gpt-oss-20b |
| MATH 500 (Vals AI) | 87.2% | 94.2% | gpt-oss-20b |
| MGSM (Vals AI) | 87.7% | 89.0% | gpt-oss-20b |
| AA Math Index | 34.7 | 89.3 | gpt-oss-20b |
| AIME 2025 (Artificial Analysis run) | 34.7% | 89.3% | gpt-oss-20b |
| OTIS Mock AIME | 38.3% | 65.3% | gpt-oss-20b |
| FrontierMath, tiers 1 to 3 | 6.0% | Not measured | Only GPT-4.1 measured |
| MATH level 5 | 83.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: math | 1369 | 1317 | GPT-4.1 |
| LMArena Text: mathematical work | 1360 | 1305 | GPT-4.1 |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index110
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| MMLU Pro | 80.5% | 71.6% | GPT-4.1 |
| LegalBench (Vals AI) | 83.1% | 70.8% | GPT-4.1 |
| CaseLaw (Vals AI) | 69.9% | 43.8% | GPT-4.1 |
| CorpFin (Vals AI) | 63.0% | 53.1% | GPT-4.1 |
| TaxEval (Vals AI) | 75.1% | 63.7% | GPT-4.1 |
| MortgageTax (Vals AI) | 65.9% | Not measured | Only GPT-4.1 measured |
| MedQA (Vals AI) | 91.2% | 82.9% | GPT-4.1 |
| MMLU Pro (Artificial Analysis run) | 80.6% | 74.8% | GPT-4.1 |
| SimpleQA Verified | 31.1% | Not measured | Only GPT-4.1 measured |
| GeoBench | 72.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: science | 1382 | 1296 | GPT-4.1 |
| LMArena Text: medicine and healthcare | 1379 | 1310 | GPT-4.1 |
| LMArena Text: legal and government | 1399 | 1280 | GPT-4.1 |
| LMArena Text: business and finance | 1369 | 1284 | GPT-4.1 |
| FACTS Benchmark Suite | 45.8% | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 3 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| Design Arena: websites | 1038 | 852 | GPT-4.1 |
| Design Arena: UI components | 997 | Not measured | Only GPT-4.1 measured |
| Design Arena: game dev | 1079 | Not measured | Only GPT-4.1 measured |
| Design Arena: data viz | 1104 | 933 | GPT-4.1 |
| Design Arena: 3D | 849 | Not measured | Only GPT-4.1 measured |
| Design Arena: all code categories | 1025 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 3 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index191
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| IFBench | 43.0% | 65.1% | gpt-oss-20b |
| Long context reasoning (AA-LCR) | 68.3% | 34.7% | GPT-4.1 |
| DTBench | 68.3% | 68.0% | GPT-4.1 |
| LMCA | 25.6% | 14.5% | GPT-4.1 |
| Fiction.LiveBench | 63.9% | Not measured | Only GPT-4.1 measured |
| LMArena Text | 1383 | 1288 | GPT-4.1 |
| LMArena Text: instruction following | 1367 | 1236 | GPT-4.1 |
| LMArena Text: creative writing | 1363 | 1201 | GPT-4.1 |
| LMArena Vision | 1210 | Not measured | Only GPT-4.1 measured |
| LMArena Text: English | 1392 | 1303 | GPT-4.1 |
| LMArena Text: non-English | 1370 | 1269 | GPT-4.1 |
| LMArena Text: Chinese | 1382 | 1315 | GPT-4.1 |
| LMArena Text: German | 1381 | 1255 | GPT-4.1 |
| LMArena Text: Japanese | 1319 | 1244 | GPT-4.1 |
| LMArena Text: Korean | 1338 | 1236 | GPT-4.1 |
| LMArena Text: French | 1382 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Spanish | 1376 | 1267 | GPT-4.1 |
| LMArena Text: Russian | 1377 | 1278 | GPT-4.1 |
| LMArena Text: Polish | 1389 | 1252 | GPT-4.1 |
| LMArena Text: multi-turn | 1397 | 1268 | GPT-4.1 |
| LMArena Text: longer queries | 1385 | 1250 | GPT-4.1 |
| LMArena Text: writing and language | 1371 | 1238 | GPT-4.1 |
| LMArena Text: entertainment and media | 1365 | 1216 | GPT-4.1 |
| LMArena Vision: English | 1209 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: Chinese | 1226 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: reading text | 1218 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: humor | 1196 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: captioning | 1207 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: recognizing things | 1208 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: creative writing | 1210 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 16 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4.1 | gpt-oss-20b | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
gpt-oss-20b costs less
Repository review
50K input + 3K output tokens
gpt-oss-20b costs less
Agent loop
200K input + 10K output tokens
gpt-oss-20b costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | OpenAI GPT-4.1 | OpenAI gpt-oss-20b |
|---|---|---|
| Maker | OpenAI | OpenAI |
| Weights | No public weights listed | Open weights |
| Released | April 14, 2025 | August 5, 2025 |
| Context window | 1.05M | 131K |
| Input $/Mtok | $2.00 | $0.03 |
| Output $/Mtok | $8.00 | $0.15 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | $0.37 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-4.1 better than gpt-oss-20b for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick GPT-4.1 for a larger context window (1.05M against 131K tokens) and work measured by LiveCodeBench (Vals AI) or LiveCodeBench (Artificial Analysis run), where it leads.
Which is cheaper, GPT-4.1 or gpt-oss-20b?
GPT-4.1 costs $2.00 per million input tokens and $8.00 per million output tokens; gpt-oss-20b costs $0.03 and $0.15. Cost per solved task is known for GPT-4.1 only.
Which has the bigger context window, GPT-4.1 or gpt-oss-20b?
GPT-4.1, with 1.05M tokens against 131K.
How many benchmarks is the GPT-4.1 vs gpt-oss-20b comparison based on?
60 benchmarks have a result for both models as of October 22, 2026. GPT-4.1's coding index rests on 0 independent benchmarks and gpt-oss-20b's on 0. Where only one model has a result, the page says so instead of naming a winner.