Compare / head-to-head
GLM-5.3-Flashvs
Hy4 preview
Hy4 preview leads 4 of 6 shared benchmarks. GLM-5.3-Flash is 5.1x cheaper per token. Both offer a 1M-token context window.
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
1 – 4
Hy4 preview leads · 1 tied
Cheaper per token
GLM-5.3-Flash
5.1x cheaper, input + output
Larger context
Tie
both 1M tokens
Providers
2 providers
Z.ai · Tencent Hunyuan
GLM-5.3-Flash
Z.ai · released 2026-08-26
textvision
Hy4 preview
Tencent Hunyuan · released 2026-08-28
text
Quality
Benchmark matrix
6 shared · 29 only GLM-5.3-Flash · 12 only Hy4 preview| Benchmark | GLM-5.3-Flash | Hy4 preview | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 6 | ||||
| GPQA Diamond | 90.2% ↗epoch run | 92.3% ↗ | -2.1 pt | Hy4 preview |
| Agents' Last Exam (ALE-CLI) | 26.3% ↗ | 22.8% ↗ | +3.5 pt | GLM-5.3-Flash |
| LMArena Agent (LMArena) | -0.0044 ↗ | 0.0395 ↗ | 0 | Tie |
| LMArena WebDev (LMArena) | 1615.6 ↗ | 1633.2 ↗ | -17.6 | Hy4 preview |
| OfficeQA Pro | 62.4% ↗ | 66.2% ↗ | -3.8 pt | Hy4 preview |
| Terminal-Bench 2.1 | 84.3% ↗ | 85.4% ↗ | -1.1 pt | Hy4 preview |
| Only GLM-5.3-Flash reports · 29 | ||||
| LMArena Elo | 1471.9 ↗ | not reported | — | — |
| APEX-Agents (Mercor) | 52.8% ↗ | not reported | — | — |
| AutomationBench v1.0.6 | 48.8% ↗ | not reported | — | — |
| BabyVision | 53.4% ↗ | not reported | — | — |
| Chartography (with tools) | 78% ↗ | not reported | — | — |
| CharXiv Reasoning (with tools) | 89.4% ↗ | not reported | — | — |
| Chess Puzzles (Epoch AI run) | 14% ↗epoch run | not reported | — | — |
| DeepSWE v1.1 | 63.4% ↗ | not reported | — | — |
| DeepSWE v1.1 (Datacurve) | 63.4% ↗ | not reported | — | — |
| FrontierMath Tier 4 v2 (Epoch AI run) | 17.1% ↗epoch run | not reported | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 55.8% ↗epoch run | not reported | — | — |
| FrontierSWE V2 (Proximal Labs) | 18.1% ↗ | not reported | — | — |
| GDPval-AA v2 | 1773 ↗ | not reported | — | — |
| HLE (with tools) | 55.3% ↗ | not reported | — | — |
| LiveBench Agentic Coding (LiveBench) | 56.8% ↗ | not reported | — | — |
| LiveBench Coding (LiveBench) | 79% ↗ | not reported | — | — |
| LiveBench Data Analysis (LiveBench) | 76.4% ↗ | not reported | — | — |
| LiveBench Instruction Following (LiveBench) | 52.8% ↗ | not reported | — | — |
| LiveBench Language (LiveBench) | 77.3% ↗ | not reported | — | — |
| LiveBench Mathematics (LiveBench) | 81.2% ↗ | not reported | — | — |
| LiveBench Reasoning (LiveBench) | 77.6% ↗ | not reported | — | — |
| LMArena Vision (LMArena) | 1278.4 ↗ | not reported | — | — |
| MMVU | 80.5% ↗ | not reported | — | — |
| MVBench | 77.8% ↗ | not reported | — | — |
| Mystery Game Puzzles (Epoch AI run) | 8% ↗epoch run | not reported | — | — |
| NL2Repo | 56.3% ↗ | not reported | — | — |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 93.9% ↗epoch run | not reported | — | — |
| Toolathlon Verified | 78.4% ↗ | not reported | — | — |
| Toolathlon-Verified (HKUST) | 78.4% ↗ | not reported | — | — |
| Only Hy4 preview reports · 12 | ||||
| CyberGym | not reported | 78.4% ↗ | — | — |
| DeepSWE | not reported | 64.3% ↗ | — | — |
| HLE (no tools, text-only) | not reported | 43.4% ↗ | — | — |
| HLE (with tools, text-only) | not reported | 55.4% ↗ | — | — |
| MathArena Apex 2025 | not reported | 74.2% ↗ | — | — |
| MCP-Atlas (public) | not reported | 83.7% ↗ | — | — |
| NL2Repo-Bench | not reported | 58.9% ↗ | — | — |
| SWE-bench Multilingual | not reported | 82.9% ↗ | — | — |
| SWE-bench Pro | not reported | 65.7% ↗ | — | — |
| Terminal-Bench 4.0 (Vals AI) | not reported | 8.1% ↗ | — | — |
| Toolathlon-Verified | not reported | 74.1% ↗ | — | — |
| WideSearch | not reported | 83.9% ↗ | — | — |
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-5.3-Flash minus Hy4 preview 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-5.3-Flash | Hy4 preview | Edge |
|---|---|---|---|
| Context window | 1M tokens | 1M tokens | Tie |
| Max output | 131K tokens | 64K tokens | GLM-5.3-Flash |
| Input price / 1M | $0.15 ↗ | $0.834 ↗ | GLM-5.3-Flash |
| Output price / 1M | $0.5 ↗ | $2.5 ↗ | GLM-5.3-Flash |
| Cached input / 1M | $0.03 ↗ | $0.042 ↗ | GLM-5.3-Flash |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $0.65 | $3.34 | GLM-5.3-Flash |
| Modalities | text · vision | text | GLM-5.3-Flash |
| Released | 2026-08-26 | 2026-08-28 | — |
| Cited benchmark scores | 40 | 18 | — |
Reliability
Provider status
Live from /statusMore matchups
GLM-5.3-Flash vs …
Models sharing the most benchmarksMore matchups
Hy4 preview vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run GLM-5.3-Flash and Hy4 preview on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.