gpt-oss-20b vs NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4
Specs, VRAM requirements and download trends — updated 9 September 2026.
openai · A · Open weights
Params
21.5 B
Context
131,072
30d
5.8 M
apache-2.0 · commercial OK
nvidia · B · Open weights
Params
44.5 B
Context
262,144
30d
193 K
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | gpt-oss-20b | NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4 |
|---|---|---|
| Deployment | Open weights · also via API | Open weights |
| Parameters | 21.5 B | 44.5 B |
| Architecture | GptOssForCausalLM | NemotronHPuzzleForCausalLM |
| Context length | 131,072 | 262,144 |
| Licence | apache-2.0 | other |
| Commercial use | Allowed | Unknown |
| Languages | — | 7 |
| Downloads 30d | 5,830,063 | 193,330 |
| Downloads all time | 108.5 M | 508 K |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | No |
| First seen on the Hub | 2025-08-04 | 2026-06-24 |
| API input price / 1M tokens | $0.03 | not on OpenRouter |
| API output price / 1M tokens | $0.13 | not on OpenRouter |
| Est. cost per 1M input + 1M output tokens | $0.16 | not on OpenRouter |
| Cached input / 1M tokens | $0.03 | not on OpenRouter |
| AA Intelligence Index | — | not on OpenRouter |
| AA Coding Index | 20.7 | not on OpenRouter |
| AA Agentic Index | 1.2 | not on OpenRouter |
| Max output tokens | 117,964 | not on OpenRouter |
| API providers | 10 | not on OpenRouter |
| Tool calling | Yes | not on OpenRouter |
| Reasoning mode | Yes | not on OpenRouter |
API cost
Estimated cost of 1M input + 1M output tokens via OpenRouter
gpt-oss-20b
$0.16
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
8.1 M193 K
gpt-oss-20b
NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
gpt-oss-20b · fp16 @ 128K ctx
82.4 GB / 24 GB
NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4 · fp16 @ 256K ctx
273.1 GB / 24 GB