head-to-head

Tencent: Hy4 flash vs Z.ai: GLM 5.3 FlashX Legal

Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-11-19.

Tencent: Hy4 flash Z.ai: GLM 5.3 FlashX Legal
Vendortencentz-ai
Quality Score100100
Input Price$0.04/M$0.37/M
Output Price$0.14/M$1.25/M
Context Window1,048,5761,048,576
Max Output65,536131,072
Tool Calling✓✓
Structured Output✓✓
Reasoning Mode✓✓
Vision✓✓
Audio--

The verdict

For a sample job of 1 million input tokens and 250,000 output tokens, Hy4 flash costs $0.070 and GLM 5.3 FlashX Legal costs $0.68, so Hy4 flash is about 9.7 times cheaper for the same work.

Both read up to 1,048,576 tokens at once, roughly 1,573 pages of text.

Only GLM 5.3 FlashX Legal offers video input.

Across the 10 tasks where they differ, Hy4 flash scores higher on 10 (SQL Generation, Code Completion, Code Documentation and others).

Pick Hy4 flash if you need lower cost or the stronger all-round task scores; pick GLM 5.3 FlashX Legal if you need video input.

Costs use current list prices per million tokens; page counts assume about 0.75 words per token and 500 words per page.

Who wins by task?

TaskTencent: Hy4 flashZ.ai: GLM 5.3 FlashX Legal
SQL Generation 134 133
Code Completion 132 131
Code Documentation 132 131
Regex Writing 120 119
CSV / Spreadsheet Cleanup 134 133
JSON Extraction 132 131
Bulk Data Labeling 129 129
Long-Document Summarization 138 137
Short-Form Summarization 124 123
Blog Post Writing 122 121

Scores reflect capability match + benchmark data + pricing for each task. Methodology →

Related comparisons

Google: Gemini 4 Neon vs Tencent: Hy4 flash Google: Gemini 4 Neon vs Z.ai: GLM 5.3 FlashX Legal Google: Gemini 4 Neon (batch) vs Tencent: Hy4 flash Google: Gemini 4 Neon (batch) vs Z.ai: GLM 5.3 FlashX Legal OpenAI: GPT-6.2 Sol Pro vs Tencent: Hy4 flash OpenAI: GPT-6.2 Sol Pro vs Z.ai: GLM 5.3 FlashX Legal OpenAI: GPT-6.2 Sol Pro (batch) vs Tencent: Hy4 flash OpenAI: GPT-6.2 Sol Pro (batch) vs Z.ai: GLM 5.3 FlashX Legal

The Model Movers Report

One email every Friday, built from this site's own rankings: the current top five by benchmark score, every model released in the last seven days, and one note worked out from that week's numbers. You can unsubscribe from any issue with one click.