SpaceXAI: Grok 4.7 vs. NVIDIA: Nemotron 3 Ultra
Direkter Head-to-Head-Vergleich zweier Frontier-Modelle. NVIDIA: Nemotron 3 Ultra gewinnt 4 von 8 Disziplinen.
Letzte Synchronisation:
|
SpaceXAI: Grok 4.7
xAI
Mehr von xAI →
|
NVIDIA: Nemotron 3 Ultra
NVIDIA
Mehr von NVIDIA →
|
|
|---|---|---|
| Quality Index | 42,2 ★ | 22,9 |
| Speed (Tokens/s) | 66,2 | 159,6 ★ |
| Latency (TTFT) | 1,76 s | 805 ms ★ |
| Preis Input (USD/1M) | $2.00 | $0.60 ★ |
| Preis Output (USD/1M) | $6.00 | $2.70 ★ |
| Context Window | — | — |
| Modalitäten | text | text |
| Release | 09/2026 | 06/2026 |
SpaceXAI: Grok 4.7
Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...
NVIDIA: Nemotron 3 Ultra
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...