Compare · ModelsLive · 2 picked · head to head
GPT-5 vs gpt-oss-20b
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-5 wins on 12/12 benchmarks
GPT-5 wins 12 of 12 shared benchmarks. Leads in knowledge · general · language.
Category leads
knowledge·GPT-5general·GPT-5language·GPT-5math·GPT-5reasoning·GPT-5coding·GPT-5
Hype vs Reality
Attention vs performance
GPT-5
#100 by perf·#4 by attention
gpt-oss-20b
#97 by perf·no signal
Best value
gpt-oss-20b
104.6x better value than GPT-5
GPT-5
9.4 pts/$
$5.63/M
gpt-oss-20b
983.3 pts/$
$0.05/M
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
12 benchmarks · 2 models
GPT-5gpt-oss-20b
Chess Puzzles
GPT-5 leads by +33.7
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GPT-5
33.7
gpt-oss-20b
0.0
Dtbench
GPT-5 leads by +37.8
GPT-5
84.5
gpt-oss-20b
46.7
GPQA diamond
GPT-5 leads by +33.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GPT-5
81.6
gpt-oss-20b
47.7
HELM · GPQA
GPT-5 leads by +19.7
GPT-5
79.1
gpt-oss-20b
59.4
HELM · IFEval
GPT-5 leads by +14.3
GPT-5
87.5
gpt-oss-20b
73.2
HELM · MMLU-Pro
GPT-5 leads by +12.3
GPT-5
86.3
gpt-oss-20b
74.0
HELM · Omni-MATH
GPT-5 leads by +8.2
GPT-5
64.7
gpt-oss-20b
56.5
HELM · WildBench
GPT-5 leads by +12.0
GPT-5
85.7
gpt-oss-20b
73.7
Lmca
GPT-5 leads by +30.0
GPT-5
47.0
gpt-oss-20b
17.1
OTIS Mock AIME 2024-2025
GPT-5 leads by +26.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GPT-5
91.4
gpt-oss-20b
65.2
Terminal Bench
GPT-5 leads by +46.2
Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence.
GPT-5
49.6
gpt-oss-20b
3.4
WeirdML
GPT-5 leads by +19.8
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GPT-5
60.7
gpt-oss-20b
40.9
Full benchmark table
| Benchmark | GPT-5 | gpt-oss-20b |
|---|---|---|
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 33.7 | 0.0 |
Dtbench | 84.5 | 46.7 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 81.6 | 47.7 |
HELM · GPQA | 79.1 | 59.4 |
HELM · IFEval | 87.5 | 73.2 |
HELM · MMLU-Pro | 86.3 | 74.0 |
HELM · Omni-MATH | 64.7 | 56.5 |
HELM · WildBench | 85.7 | 73.7 |
Lmca | 47.0 | 17.1 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 91.4 | 65.2 |
Terminal Bench Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence. | 49.6 | 3.4 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 60.7 | 40.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $1.25 | $10.00 | 400K tokens (~200 books) | $34.38 | |
| $0.02 | $0.09 | 131K tokens (~66 books) | $0.36 |