Compare / head-to-head
GLM-5.2vsNVIDIA Nemotron 3 Ultra 550B-A55B NVFP4
GLM-5.2 leads 2 of 2 shared benchmarks. Both offer a 1M-token context window.
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
2 – 0
GLM-5.2 leads
Cheaper per token
—
list price, input + output
Larger context
Tie
both 1M tokens
Provider uptime (30d)
— · 100%
Z.ai · NVIDIA
GLM-5.2
Z.ai · released 2026-06-24
text
NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4
NVIDIA · released 2026-06-04
text
- Context
- 1M
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 6 · 4 core
Quality
Benchmark matrix
2 shared · 23 only GLM-5.2 · 4 only NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4| Benchmark | GLM-5.2 | NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 2 | ||||
| GPQA Diamond | 91.9% ↗epoch | 87.9% ↗ | +4 pt | GLM-5.2 |
| Terminal-Bench 2.1 | 81% ↗ | 53.9% ↗ | +27.1 pt | GLM-5.2 |
| Only GLM-5.2 reports · 23 | ||||
| AIME 2026 | 90% ↗matharena ⚠ | not reported | — | — |
| Agents' Last Exam (ALE-CLI) | 23.8% ↗ | not reported | — | — |
| AutomationBench v1.0.6 | 26.2% ↗ | not reported | — | — |
| CyberGym | 77.2% ↗ | not reported | — | — |
| DeepSWE v1.1 | 46.2% ↗ | not reported | — | — |
| ExploitBench | 24.4% ↗ | not reported | — | — |
| ExploitGym 2h | 29 ↗ | not reported | — | — |
| ExploitGym 6h | 39 ↗ | not reported | — | — |
| FrontierMath Tier 4 v2 (Epoch AI run) | 29.3% ↗epoch | not reported | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 59.2% ↗epoch | not reported | — | — |
| FrontierSWE | 67.5% ↗ | not reported | — | — |
| GDPval-AA v2 | 1508 ↗ | not reported | — | — |
| HLE (with tools) | 54.7% ↗ | not reported | — | — |
| NL2Repo | 48.9% ↗ | not reported | — | — |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 86.4% ↗epoch | not reported | — | — |
| PostTrainBench | 31.7% ↗ | not reported | — | — |
| ProgramBench (Almost Solved) | 9.5% ↗ | not reported | — | — |
| SimpleQA Verified | 34.2% ↗epoch | not reported | — | — |
| SWE-Bench Pro | 62.1% ↗ | not reported | — | — |
| SWE-bench Verified (Epoch AI run) | 78.7% ↗epoch | not reported | — | — |
| SWE-Marathon v1.1 | 19.4% ↗ | not reported | — | — |
| Terminal-Bench 3.0 | 4.6% ↗ | not reported | — | — |
| Toolathlon Verified | 59.9% ↗ | not reported | — | — |
| Only NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 reports · 4 | ||||
| SWE-bench Verified | not reported | 69.7% ↗ | — | — |
| Humanity's Last Exam (no tools) | not reported | 26.1% ↗ | — | — |
| LMArena Elo | not reported | 1444.7 ↗ | — | — |
| BrowseComp | not reported | 41.4% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is GLM-5.2 minus NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | GLM-5.2 | NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 | Edge |
|---|---|---|---|
| Context window | 1M tokens | 1M tokens | Tie |
| Max output | 131K tokens | — | — |
| Input price / 1M | $1.4 ↗ | — | — |
| Output price / 1M | $4.4 ↗ | — | — |
| Cached input / 1M | $0.26 ↗ | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $5.8 | — | — |
| Modalities | text | text | Tie |
| Released | 2026-06-24 | 2026-06-04 | — |
| Cited benchmark scores | 26 | 6 | — |
Reliability
Provider status
Last 60 days · refreshed daily on this page · live on /statusMore matchups
GLM-5.2 vs …
Models sharing the most benchmarksMore matchups
NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run GLM-5.2 and NVIDIA Nemotron 3 Ultra 550B-A55B NVFP4 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.