Enter opens the highlighted result. Esc closes.

GPT-4 vs gpt-oss-20b

Updated October 21, 2026. 34 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-4 Replace gpt-oss-20b
GPT-4OpenAI
Coding
No data
Overall
6.7
Output $/M
$60.00
2
gpt-oss-20bOpenAI
Coding
No data
Overall
10.0
Output $/M
$0.15
32
vs boards won of 34 shared

Decision reading

Neither model has a coding score yet, so the two cannot be ranked on coding.

GPT-4 has the edge where what matters is LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model AGPT-4OpenAINot benchmarkedNo coding benchmarkNo cost per solved taskModel Bgpt-oss-20bOpenAINot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
34
GPT-4 only
11
gpt-oss-20b only
26
Like-for-like categories
6 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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.

At a glance

Coding index
GPT-4 not benchmarked, gpt-oss-20b not benchmarked.
Widest category
Knowledge, 20.8 points apart on average over 5 shared benchmarks.
Widest single gap
GPQA Diamond (Artificial Analysis run): GPT-4 34.9% against gpt-oss-20b 68.8%, 33.9 points apart on a 0 to 100 scale.
Output price
GPT-4 $60.00 against gpt-oss-20b $0.15 per million output tokens, 400.0x apart.
Context window
GPT-4 takes 8K 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.

Boards won, where both have a result
CategoryGPT-4Barsgpt-oss-20bTies
Coding02
Intelligence02
Reasoning04
Math02
Knowledge05
General217
All shared boards232
CodingLike-for-likeSix of these boards feed the Vibe Coding Index02
Coding benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
SWE-bench Verified22.4%Not measuredOnly GPT-4 measured
SWE-bench Lite39.7%Not measuredOnly GPT-4 measured
SWE-bench (full test set)12.5%Not measuredOnly GPT-4 measured
AA Coding Index13.1Not measuredOnly GPT-4 measured
LiveCodeBench (Vals AI)Not measured80.4%Only gpt-oss-20b measured
LiveCodeBench (Artificial Analysis run)Not measured77.7%Only gpt-oss-20b measured
SciCodeNot measured38.9%Only gpt-oss-20b measured
Terminal-Bench Hard (Artificial Analysis run)Not measured10.6%Only gpt-oss-20b measured
Terminal-BenchNot measured3.4%Only gpt-oss-20b measured
WeirdMLNot measured40.9%Only gpt-oss-20b measured
ALE-BenchNot measured566.0Only gpt-oss-20b measured
LMArena Text: coding12091307gpt-oss-20b
LMArena Text: software and IT11951311gpt-oss-20b
Neither model has a result on the other 25 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
AA Intelligence Index6.710.0gpt-oss-20b
Epoch Capabilities Index125.7137.7gpt-oss-20b
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
Agentic work benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
τ²-Bench TelecomNot measured60.2%Only gpt-oss-20b measured
Neither model has a result on the other 38 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index04
Reasoning benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
GPQA DiamondNot measured68.9%Only gpt-oss-20b measured
GPQA Diamond (Artificial Analysis run)34.9%68.8%gpt-oss-20b
Humanity's Last Exam (Artificial Analysis run)Not measured11.0%Only gpt-oss-20b measured
GPQA Diamond (Epoch AI run)Not measured60.8%Only gpt-oss-20b measured
BIG-Bench Hard75.1%Not measuredOnly GPT-4 measured
LMArena Text: hard prompts12001274gpt-oss-20b
LMArena Text: expert11491258gpt-oss-20b
LMArena Text: hard prompts in English12091292gpt-oss-20b
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index02
Math benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
AIMENot measured86.0%Only gpt-oss-20b measured
MATH 500 (Vals AI)Not measured94.2%Only gpt-oss-20b measured
MGSM (Vals AI)Not measured89.0%Only gpt-oss-20b measured
AA Math IndexNot measured89.3Only gpt-oss-20b measured
AIME 2025 (Artificial Analysis run)Not measured89.3%Only gpt-oss-20b measured
OTIS Mock AIMENot measured65.3%Only gpt-oss-20b measured
GSM8K92.0%Not measuredOnly GPT-4 measured
LMArena Text: math12301317gpt-oss-20b
LMArena Text: mathematical work12281305gpt-oss-20b
Neither model has a result on the other 4 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index05
Knowledge benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
MMLU ProNot measured71.6%Only gpt-oss-20b measured
LegalBench (Vals AI)Not measured70.8%Only gpt-oss-20b measured
CaseLaw (Vals AI)Not measured43.8%Only gpt-oss-20b measured
CorpFin (Vals AI)Not measured53.1%Only gpt-oss-20b measured
TaxEval (Vals AI)Not measured63.7%Only gpt-oss-20b measured
MedQA (Vals AI)Not measured82.9%Only gpt-oss-20b measured
MMLU Pro (Artificial Analysis run)56.2%74.8%gpt-oss-20b
MMLU86.4%Not measuredOnly GPT-4 measured
AI2 Reasoning Challenge96.3%Not measuredOnly GPT-4 measured
LMArena Text: science11771296gpt-oss-20b
LMArena Text: medicine and healthcare11451310gpt-oss-20b
LMArena Text: legal and government11911280gpt-oss-20b
LMArena Text: business and finance11451284gpt-oss-20b
Neither model has a result on the other 4 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
Web design benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
Design Arena: websitesNot measured853Only gpt-oss-20b measured
Design Arena: data vizNot measured933Only 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 index217
General benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
IFBench33.2%65.1%gpt-oss-20b
Long context reasoning (AA-LCR)Not measured34.7%Only gpt-oss-20b measured
DTBench62.7%68.0%gpt-oss-20b
LMCA17.1%14.5%GPT-4
HellaSwag95.3%Not measuredOnly GPT-4 measured
WinoGrande87.5%Not measuredOnly GPT-4 measured
LMArena Text12061288gpt-oss-20b
LMArena Text: instruction following12011236gpt-oss-20b
LMArena Text: creative writing11911201gpt-oss-20b
LMArena Text: English12201303gpt-oss-20b
LMArena Text: non-English11941269gpt-oss-20b
LMArena Text: Chinese11841315gpt-oss-20b
LMArena Text: German11981255gpt-oss-20b
LMArena Text: Japanese11371244gpt-oss-20b
LMArena Text: Korean10871236gpt-oss-20b
LMArena Text: French1220Not measuredOnly GPT-4 measured
LMArena Text: Spanish11971267gpt-oss-20b
LMArena Text: Russian11951278gpt-oss-20b
LMArena Text: Polish12651252GPT-4
LMArena Text: multi-turn12061268gpt-oss-20b
LMArena Text: longer queries11901250gpt-oss-20b
LMArena Text: writing and language12191238gpt-oss-20b
LMArena Text: entertainment and media11841216gpt-oss-20b
Neither model has a result on the other 23 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-4 and gpt-oss-20b
BenchmarkGPT-4gpt-oss-20bLeader
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-4$0.060Fits in one request
gpt-oss-20b$0.0001Fits in one request

gpt-oss-20b costs less

Repository review

50K input + 3K output tokens

GPT-4$1.68Exceeds the context window
gpt-oss-20b$0.0019Fits in one request

gpt-oss-20b costs less

Agent loop

200K input + 10K output tokens

GPT-4$6.60Exceeds the context window
gpt-oss-20b$0.0075Exceeds the context window

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.

Specifications of GPT-4 and gpt-oss-20b
SpecificationOpenAI GPT-4OpenAI gpt-oss-20b
MakerOpenAIOpenAI
WeightsNo public weights listedOpen weights
ReleasedMay 28, 2023August 5, 2025
Context window8K131K
Input $/Mtok$30.00$0.03
Output $/Mtok$60.00$0.15
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-4 better than gpt-oss-20b for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. GPT-4 has the edge where what matters is LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-4 or gpt-oss-20b?

GPT-4 costs $30.00 per million input tokens and $60.00 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-4 or gpt-oss-20b?

gpt-oss-20b, with 131K tokens against 8K.

How many benchmarks is the GPT-4 vs gpt-oss-20b comparison based on?

34 benchmarks have a result for both models as of October 21, 2026. GPT-4'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.

Compare next