GPT-4o (2024-08-06) vs gpt-oss-20b
gpt-oss-20b leads GPT-4o (2024-08-06) on the Artificial Analysis Intelligence Index, 10.0 to 7.7, and gpt-oss-20b costs 66.7x less per output token ($0.15 vs $10.00 per million). Updated October 22, 2026; 44 benchmarks have a result for both.
GPT-4o (2024-08-06) or gpt-oss-20b: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick GPT-4o (2024-08-06) for work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.
- Shared results
- 44
- GPT-4o (2024-08-06) only
- 14
- gpt-oss-20b only
- 19
- Like-for-like categories
- 7 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-oss-20b
gpt-oss-20b leads, 10.0 against 7.7.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only gpt-oss-20b has a GPQA Diamond result, 68.9%.
Directional only
Math
AIME
gpt-oss-20b
gpt-oss-20b leads, 86.0% against 14.0%.
Same board, both measured
Knowledge
MMLU Pro
GPT-4o (2024-08-06)
GPT-4o (2024-08-06) leads, 74.1% against 71.6%.
Same board, both measured
Web design
Design Arena: websites
No call
Only gpt-oss-20b has a Design Arena: websites result, 852.
Directional only
Long documents
Listed context window
gpt-oss-20b
gpt-oss-20b takes 131K tokens against 128K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
gpt-oss-20b
GPT-4o (2024-08-06) $0.0075, gpt-oss-20b $0.0001 per turn.
List prices
At a glance
- Coding index
- GPT-4o (2024-08-06) not benchmarked, gpt-oss-20b not benchmarked.
- Widest category
- Agentic work, 31.3 points apart on average over 1 shared benchmark.
- Widest single gap
- AIME: GPT-4o (2024-08-06) 14.0% against gpt-oss-20b 86.0%, 72.1 points apart on a 0 to 100 scale.
- Output price
- GPT-4o (2024-08-06) $10.00 against gpt-oss-20b $0.15 per million output tokens, 66.7x apart.
- Context window
- GPT-4o (2024-08-06) takes 128K 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-4o (2024-08-06) | Bars | gpt-oss-20b | Ties |
|---|---|---|---|---|
| Coding | 0 | 4 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 0 | 6 | ||
| Math | 1 | 5 | ||
| Knowledge | 4 | 5 | ||
| General | 9 | 8 | ||
| All shared boards | 14 | 30 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| Aider polyglot | 23.1% | Not measured | Only GPT-4o (2024-08-06) measured |
| SWE-bench Lite | 21.7% | Not measured | Only GPT-4o (2024-08-06) measured |
| LiveCodeBench (Vals AI) | Not measured | 80.4% | Only gpt-oss-20b measured |
| LiveCodeBench (Artificial Analysis run) | 31.7% | 77.7% | gpt-oss-20b |
| SciCode | Not measured | 38.9% | Only gpt-oss-20b measured |
| Terminal-Bench Hard (Artificial Analysis run) | 8.3% | 10.6% | gpt-oss-20b |
| Terminal-Bench | Not measured | 3.4% | Only gpt-oss-20b measured |
| WeirdML | Not measured | 40.9% | Only gpt-oss-20b measured |
| ALE-Bench | Not measured | 566.0 | Only gpt-oss-20b measured |
| LMArena Text: coding | 1283 | 1307 | gpt-oss-20b |
| LMArena Text: software and IT | 1278 | 1311 | gpt-oss-20b |
| Neither model has a result on the other 32 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| AA Intelligence Index | 7.7 | 10.0 | gpt-oss-20b |
| Epoch Capabilities Index | Not measured | 137.7 | Only gpt-oss-20b measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 28.9% | 60.2% | gpt-oss-20b |
| BrowseComp-Plus | Not measured | 50.6% | Only gpt-oss-20b measured |
| Neither model has a result on the other 40 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index06
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 68.9% | Only gpt-oss-20b measured |
| MMMU Pro (Vals AI) | 64.0% | Not measured | Only GPT-4o (2024-08-06) measured |
| GPQA Diamond (Artificial Analysis run) | 52.1% | 68.8% | gpt-oss-20b |
| Humanity's Last Exam (Artificial Analysis run) | 2.3% | 11.0% | gpt-oss-20b |
| GPQA Diamond (Epoch AI run) | 49.2% | 60.8% | gpt-oss-20b |
| SimpleBench | 17.8% | Not measured | Only GPT-4o (2024-08-06) measured |
| Chess puzzles | 6.0% | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Text: hard prompts | 1264 | 1274 | gpt-oss-20b |
| LMArena Text: expert | 1241 | 1258 | gpt-oss-20b |
| LMArena Text: hard prompts in English | 1277 | 1292 | gpt-oss-20b |
| Kagi LLM Benchmark | Not measured | 53.2% | Only gpt-oss-20b measured |
| Neither model has a result on the other 12 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index15
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| AIME | 14.0% | 86.0% | gpt-oss-20b |
| MATH 500 (Vals AI) | 75.2% | 94.2% | gpt-oss-20b |
| MGSM (Vals AI) | 90.7% | 89.0% | GPT-4o (2024-08-06) |
| AA Math Index | Not measured | 89.3 | Only gpt-oss-20b measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 89.3% | Only gpt-oss-20b measured |
| OTIS Mock AIME | 6.4% | 65.3% | gpt-oss-20b |
| FrontierMath, tiers 1 to 3 | 0.4% | Not measured | Only GPT-4o (2024-08-06) measured |
| MATH level 5 | 53.3% | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Text: math | 1285 | 1317 | gpt-oss-20b |
| LMArena Text: mathematical work | 1283 | 1305 | gpt-oss-20b |
| MathArena | Not measured | 12.8% | Only gpt-oss-20b measured |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index45
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| MMLU Pro | 74.1% | 71.6% | GPT-4o (2024-08-06) |
| LegalBench (Vals AI) | 80.1% | 70.8% | GPT-4o (2024-08-06) |
| CaseLaw (Vals AI) | Not measured | 43.8% | Only gpt-oss-20b measured |
| CorpFin (Vals AI) | 39.4% | 53.1% | gpt-oss-20b |
| TaxEval (Vals AI) | 71.1% | 63.7% | GPT-4o (2024-08-06) |
| MortgageTax (Vals AI) | 61.0% | Not measured | Only GPT-4o (2024-08-06) measured |
| MedQA (Vals AI) | 88.2% | 82.9% | GPT-4o (2024-08-06) |
| MMLU Pro (Artificial Analysis run) | Not measured | 74.8% | Only gpt-oss-20b measured |
| SimpleQA Verified | 26.0% | Not measured | Only GPT-4o (2024-08-06) measured |
| MMLU | 84.3% | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Text: science | 1274 | 1296 | gpt-oss-20b |
| LMArena Text: medicine and healthcare | 1252 | 1310 | gpt-oss-20b |
| LMArena Text: legal and government | 1275 | 1280 | gpt-oss-20b |
| LMArena Text: business and finance | 1232 | 1284 | gpt-oss-20b |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 852 | Only gpt-oss-20b measured |
| Design Arena: data viz | Not measured | 933 | Only gpt-oss-20b measured |
| Neither model has a result on the other 7 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index98
| Benchmark | GPT-4o (2024-08-06) | gpt-oss-20b | Leader |
|---|---|---|---|
| IFBench | 36.0% | 65.1% | gpt-oss-20b |
| Long context reasoning (AA-LCR) | 41.0% | 34.7% | GPT-4o (2024-08-06) |
| DTBench | Not measured | 68.0% | Only gpt-oss-20b measured |
| LMCA | Not measured | 14.5% | Only gpt-oss-20b measured |
| LMArena Text | 1283 | 1288 | gpt-oss-20b |
| LMArena Text: instruction following | 1267 | 1236 | GPT-4o (2024-08-06) |
| LMArena Text: creative writing | 1274 | 1201 | GPT-4o (2024-08-06) |
| LMArena Vision | 1065 | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Text: English | 1297 | 1303 | gpt-oss-20b |
| LMArena Text: non-English | 1262 | 1269 | gpt-oss-20b |
| LMArena Text: Chinese | 1254 | 1315 | gpt-oss-20b |
| LMArena Text: German | 1257 | 1255 | GPT-4o (2024-08-06) |
| LMArena Text: Japanese | 1234 | 1244 | gpt-oss-20b |
| LMArena Text: Korean | 1218 | 1236 | gpt-oss-20b |
| LMArena Text: French | 1263 | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Text: Spanish | 1269 | 1267 | GPT-4o (2024-08-06) |
| LMArena Text: Russian | 1272 | 1278 | gpt-oss-20b |
| LMArena Text: Polish | Not measured | 1252 | Only gpt-oss-20b measured |
| LMArena Text: multi-turn | 1279 | 1268 | GPT-4o (2024-08-06) |
| LMArena Text: longer queries | 1283 | 1250 | GPT-4o (2024-08-06) |
| LMArena Text: writing and language | 1290 | 1238 | GPT-4o (2024-08-06) |
| LMArena Text: entertainment and media | 1267 | 1216 | GPT-4o (2024-08-06) |
| LMArena Vision: English | 1072 | Not measured | Only GPT-4o (2024-08-06) measured |
| LMArena Vision: Chinese | 1009 | Not measured | Only GPT-4o (2024-08-06) measured |
| Neither model has a result on the other 24 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4o (2024-08-06) | 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-4o (2024-08-06) | OpenAI gpt-oss-20b |
|---|---|---|
| Maker | OpenAI | OpenAI |
| Weights | No public weights listed | Open weights |
| Released | August 6, 2024 | August 5, 2025 |
| Context window | 128K | 131K |
| Input $/Mtok | $2.50 | $0.03 |
| Output $/Mtok | $10.00 | $0.15 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-4o (2024-08-06) better than gpt-oss-20b for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick GPT-4o (2024-08-06) for work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.
Which is cheaper, GPT-4o (2024-08-06) or gpt-oss-20b?
GPT-4o (2024-08-06) costs $2.50 per million input tokens and $10.00 per million output tokens; gpt-oss-20b costs $0.03 and $0.15. Cost per solved task is known for neither of them.
Which has the bigger context window, GPT-4o (2024-08-06) or gpt-oss-20b?
gpt-oss-20b, with 131K tokens against 128K.
How many benchmarks is the GPT-4o (2024-08-06) vs gpt-oss-20b comparison based on?
44 benchmarks have a result for both models as of October 22, 2026. GPT-4o (2024-08-06)'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.