Model
Nemotron 3 Ultra 550B A55B
Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy
Compare providers (sorted by listed input price)
12 of 15 providers publish a price for Nemotron 3 Ultra 550B A55B. The cheapest full-context option is DevPass (LLM Gateway) at $0.5/$2.2 per M — 9% below Nvidia's own price. routing.run lists lower but caps context at 131K.
Best price history
Daily lowest published price across live providers. A lower line is better for the buyer.
- Best input /M
- $0.1unchanged
- Best output /M
- $0.1unchanged
- Provider coverage
- 16
- 16 at the start · 24 snapshots
Published best prices are unchanged.
24 daily snapshots collected from Aug 21, 2026 through Sep 14, 2026. A price chart will appear after the first real move.
Price badge
https://www.vaanalytics.in/badge/nvidia/nemotron-3-ultra-550b-a55b.svgLive SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.
Recent changes
Changelog →- contextvia kilo · Sep 14, 2026
- context: 256K → 203K tokens
- output: 33K → 183K tokens
- repricedvia openrouter · Sep 14, 2026
- output: 33K → 183K tokens
- input: $0.625 → $0.6
- output: $3.13 → $2.40
- cache_read: $0.188 → $0.12
- capabilityvia edenai · Sep 7, 2026
- name: Nemotron 3 Ultra 550B A55B → Nemotron 3 Ultra 550B A55B (Nebius)
- capabilityvia edenai · Sep 7, 2026
- name: Nemotron 3 Ultra 550B A55B → Nemotron 3 Ultra 550B A55B (Deep Infra)
- repricedvia openrouter · Sep 4, 2026
- output: 183K → 33K tokens
- input: $0.6 → $0.625
- output: $2.40 → $3.13
- cache_read: $0.12 → $0.188
- contextvia kilo · Sep 4, 2026
- context: 203K → 256K tokens
- output: 183K → 33K tokens
Benchmarks
- SWE-Bench Verified ↗(resolved)report ↗Unclassified source70.7
- SWE-Bench Multilingual ↗(resolve rate)report ↗Unclassified source67.7
- Terminal-Bench ↗(success rate)report ↗Unclassified source56.4
- GPQA ↗(accuracy)report ↗Unclassified source87
- Humanity's Last Exam ↗(accuracy)report ↗Unclassified source26.7
- Humanity's Last Exam ↗(accuracy)report ↗Unclassified source37.4
- LiveCodeBench ↗(pass@1)report ↗Unclassified source89
- MMLU-Pro ↗(accuracy)report ↗Unclassified source86.8
- BrowseComp ↗(accuracy)report ↗Unclassified source44.4
- IFBench(accuracy)report ↗Unclassified source81.7
- GDPval(wins or ties)report ↗Unclassified source46.7
Scores a lab published about its own model are claims, not measurements. Rankings elsewhere on this site use independently measured scores only.
Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.