Manufacturer

OpenAI

Offene gpt-oss-Modelle von OpenAI.

2Models

190Benchmarks

190Performance

0Harness

Best token generation

1.995,4 tok/s

gpt-oss-20b

Best prefill

35.347,7 tok/s

gpt-oss-20b

∅ Token generation

398,0 tok/s

Average of 190 runs

∅ Prefill

5.917,5 tok/s

Average of 190 runs

Best TTFT

114 ms

gpt-oss-20b

Best harness rate

no harness run yet

🏆

gpt-oss-20b · fastest model at 1.995,4 tok/s token generation
Best prefill of this manufacturer: 35.347,7 tok/s (gpt-oss-20b)

Best measured generation throughput per model (tok/s), published performance benchmarks.

gpt-oss-20b

1.995,4 tok/s

gpt-oss-120b

1.335,9 tok/s

Models

2 assigned