Gemini 3 Pro vs GLM-5
Benchmarks, API pricing and specs, head to head. Data updated 2026-06-10.
Google · Nov 2025
Google's November 2025 frontier breakout — 91.9% GPQA Diamond and 37.5% HLE made it the reasoning leader of its generation.
Z.ai (Zhipu) · Feb 2026
Z.ai's 744B open-weights MoE — within ~4 points of Opus 4.6 on SWE-bench Verified at $1/$3.20 per million tokens.
The verdict
GLM-5 wins 1 of the 1 benchmarks these models share, against 0 for Gemini 3 Pro. GLM-5 is about 2.9x cheaper per blended million tokens (3:1 input:output mix). Gemini 3 Pro also takes 1M of context versus 200K for GLM-5. And GLM-5 is open-weights (MIT), so it can be self-hosted — a structural advantage if data control or vendor independence matters.
Benchmark head-to-head 0–1
Specs & pricing
| Gemini 3 Pro | GLM-5 | |
|---|---|---|
| modhub Index | 73.6 | — |
| Input price / 1M | $2 | $1 |
| Output price / 1M | $12 | $3.2 |
| Context window | 1M | 200K |
| Max output | 66K | 128K |
| Open weights | no | yes (MIT) |
| Reasoning model | yes | yes |
| Multimodal input | text, image, audio, video | text |
| Knowledge cutoff | Jan 2025 | Dec 2025 |
| Released | Nov 2025 | Feb 2026 |
| Example monthly cost* | $38.00 | $14.80 |
* 10M input + 1.5M output tokens per month at list prices, no caching. Green = better value on that row.
Frequently asked questions
- Which is better, Gemini 3 Pro or GLM-5?
- GLM-5 wins 1 of the 1 benchmarks these models share, against 0 for Gemini 3 Pro. GLM-5 is about 2.9x cheaper per blended million tokens (3:1 input:output mix). Gemini 3 Pro also takes 1M of context versus 200K for GLM-5. And GLM-5 is open-weights (MIT), so it can be self-hosted — a structural advantage if data control or vendor independence matters.
- Which is cheaper, Gemini 3 Pro or GLM-5?
- Gemini 3 Pro costs $2/$12 per million input/output tokens, while GLM-5 costs $1/$3.2. For a typical workload of 10M input and 1.5M output tokens per month, that's $38.00 versus $14.80.
- Which model is better for coding, Gemini 3 Pro or GLM-5?
- On SWE-bench Verified — the standard agentic-coding benchmark — GLM-5 scores ~77.8% versus 76.2% for Gemini 3 Pro, making GLM-5 the stronger pick for coding agents.