- Score
- 8.6 (+0.58 points)
- Total Cost (Current Price)
- $0.061 · 91.0% cheaper
- Response Time (avg)
- 7.51s · 81.4% faster
NVIDIA
Release: 2026-06-04
Tested on: 2026-10-01 02:17
nvidia/nemotron-3-ultra-550b-a55b::medium
550B total (55B active)MoEWeights available
(medium)
(none)
Summary
Nemotron 3 Ultra scores 8.0 on AI BENCHY and ranks #109. It has 9.8 reliability, a 68.1% pass rate, $0.674 total cost, and 40.33s average response time.
What makes Nemotron 3 Ultra unique: It stands out most in Coding, where it ranks #3, while Combined is its weakest area at #17.
Recommended alternatives
We recommend choosing one of these models instead:
- Score
- 8.3 (+0.27 points)
- Total Cost (Current Price)
- $0.047 · 93.0% cheaper
- Response Time (avg)
- 18.07s · 55.2% faster
- Score
- 9.4 (+1.37 points)
- Total Cost (Current Price)
- $0.183 · 73.0% cheaper
- Response Time (avg)
- 6.24s · 84.5% faster
Model facts
Researched on 2026-08-12
- Parameters
- 550B total (55B active)
- Architecture
- MoE
- Availability
- Weights available
- License
- NVIDIA Open Model
8.0
Consistency
8.9
9.8
Total Output Tokens
196,626
Total Input Tokens
482,754
Input Price
$0.500 / 1M
Output Price
$2.200 / 1M
Cache Read Price
$0.100 / 1M
Cache Write Price
N/A
Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.
Flaky tests
3
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#109 Nemotron 3 Ultra
medium
Reached the allocated time limit (300 seconds) without receiving showcase output.
- Cost
- $0.000
- Time
- 300.0s
- Tokens
- 0 tok
Run history
| Tested on | Score | Reliability | Tests Correct | Total Cost | Compare |
|---|---|---|---|---|---|
| 2026-10-01 02:17 Re-test | 8.0 | 9.8 | $0.674 |