LFM2.5-8B-A1B vs Nemotron 3 Ultra 550B A55B: Benchmarks, Pricing & Capabilities Compared
TL;DR — LFM2.5-8B-A1B wins for general use · Nemotron 3 Ultra 550B A55B wins for reasoning + long-context.
LFM2.5-8B-A1B Liquid AI
- Released
- 2026-05-28
- Context window
- 32K tokens
- Input price
- $0.00 / Mtok
- Output price
- $0.00 / Mtok
Nemotron 3 Ultra 550B A55B NVIDIA
- Released
- 2026-06-04
- Context window
- 262K tokens
- Input price
- $0.60 / Mtok
- Output price
- $2.60 / Mtok
Benchmark comparison
| Benchmark | LFM2.5-8B-A1B | Nemotron 3 Ultra 550B A55B |
|---|---|---|
| AA Intelligence Index | 14.2 | 47.7 ✓ |
| GPQA Diamond | 51.3% | 86.7% ✓ |
| HLE | 6.9% | 26.6% ✓ |
| IF-Bench | 55.6% | 81.4% ✓ |
| LiveCodeBench Reasoning | 0.0% | 67.0% ✓ |
| SciCode | 7.8% | 39.9% ✓ |
| TAU2-bench | 16.1% | 83.3% ✓ |
| TerminalBench-Hard | 4.5% | 36.4% ✓ |
Pricing comparison
| Metric | LFM2.5-8B-A1B | Nemotron 3 Ultra 550B A55B |
|---|---|---|
| Input ($/Mtok) | $0.00 | $0.60 |
| Output ($/Mtok) | $0.00 | $2.60 |
| Cached input ($/Mtok) | — | — |
| Cost per 1M-token roundtrip (1M in + 1M out) | $0.00 | $3.20 |
Context window & modalities
| Attribute | LFM2.5-8B-A1B | Nemotron 3 Ultra 550B A55B |
|---|---|---|
| Context window | 32K tokens | 262K tokens |
| Input modalities | text | text |
| Output modalities | text | text |
| Knowledge cutoff | — | — |
Verdict by use case
Coding
Insufficient data
Basis: SWE-bench
No shared coding benchmark.
Reasoning
→ Nemotron 3 Ultra 550B A55B
Basis: GPQA Diamond
LFM2.5-8B-A1B 51.3% vs Nemotron 3 Ultra 550B A55B 86.7% on GPQA Diamond.
Math
Insufficient data
Basis: MATH / AIME
No shared math benchmark.
Long context
→ Nemotron 3 Ultra 550B A55B
Basis: Context window
LFM2.5-8B-A1B 32K tokens vs Nemotron 3 Ultra 550B A55B 262K tokens.
Cost
→ LFM2.5-8B-A1B
Basis: Input $/Mtok
LFM2.5-8B-A1B $0/Mtok vs Nemotron 3 Ultra 550B A55B $0.6/Mtok input.
Changelog & releases
LFM2.5-8B-A1B
Released 2026-05-28
Nemotron 3 Ultra 550B A55B
Released 2026-06-04