• 3 % discount on prepayment
  • Insured shipping from Switzerland
  • Sustainable: refurbished instead of new
  • Warranty on all cards
← Back to overview

gpt-oss-120b

OpenAI
Apache 2.0

Open language model for chat and text. 2 sizes from 20.9B to 116.8B parameters, partly mixture-of-experts (only a part of the parameters is active per token).

  • Mixture-of-experts: 116.8B parameters in total, of which 5.8B active per token. Memory requirement depends on all parameters, speed on the active ones.

Settings

Quality
4 bit (Q4_K_M) is the most common choice: good quality with low memory requirements.
Context length: 8k tokens
4k128k
How much text the model keeps in view at once. More context needs more memory.
Recommendation
1× H100 80 GB

The model needs about 71.9 GB of memory. Hopper, FP8 support.

Weights
70.1 GB
KV cache
0.3 GB
Speed
~281.6 Tok./s

More suitable configurations

CardQuantityTok./sStatus
Tesla T4 16 GB8×31.22-3 weeks delivery time
Tesla P100 16 GB8×39.32-3 weeks delivery time
Tesla P40 24 GB4×46.42-3 weeks delivery time
Tesla V100 16 GB8×87.0only 1 in stock
Tesla V100 32 GB4×109.4On order, delivery time approx. 1 week
AMD Instinct MI50 16 GB8×66.7only 2 in stock
AMD Instinct MI50 32 GB4×83.9only 2 in stock
AMD Instinct MI60 32 GB4×77.52-3 weeks delivery time
AMD Instinct MI100 32 GB4×93.72-3 weeks delivery time
A30 24 GB4×105.92-3 weeks delivery time
RTX A5000 24 GB4×110.32-3 weeks delivery time
A40 48 GB2×120.82-3 weeks delivery time