GPT-OSS 120B vs GPT-OSS 20B
Change modelsGPT-OSS 120BOpenAI
GPT-OSS 20BOpenAI
IndexAll categories, out of 100.
43.4 ±4.2 best
41.7 ±4.6
GapIndex points below the top.
Top
−1.7 · no clear gap
CoverageIndex weight that has results.
90%
90%
SpeedOutput tokens per second.
34/s
310/s best
EloLMArena rating and rank.
1365 best
1288
ReleasedMonth of release.
2025-08
2025-08
Input / 1MUS dollars per 1M input tokens.
$0.15
$0.03 best
Output / 1MUS dollars per 1M output tokens.
$0.60
$0.13 best
1K chats2K tokens in, 500 out, each.
$0.60
$0.125 best
1K documents20K tokens in, 1K out, each.
$3.60
$0.73 best
1K agent tasks200K tokens in, 20K out, each.
$42.00
$8.60 best
ContextMaximum tokens in one request.
131K
131K
AgenticMulti-step tasks with tools.
40.1 best
39.8
CodingCode writing and repair.
42.8
48.9 best
ReasoningLogic problems and puzzles.
40.2 best
35.3
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
42.4 best
36.0
MultilingualTasks in many languages.
51.1 best
37.0
InstructionTasks with strict prompt rules.
53.5 best
49.9
MathMath problems.
51.5 best
45.9
Benchmarks
37 sharedOnly GPT-OSS 120B: 11
GPT-OSS 120B
GPT-OSS 20B