Gemini 4 Carbon and GLM 5.3 compared
Gemini 4 Carbon, GLM 5.3, compared side by side on price, context, and capabilities. All are available on Hormouz through one OpenAI-compatible API.
Google Gemini 4 Carbon | Z.ai GLM 5.3 | |
|---|---|---|
| At a glance | ||
| Context window | 1,048,576 | 1,048,576 |
| Input type | audio, file, image, text, video | text |
| Output type | text | text |
| Type | text, audio, image, video | text |
| Pricing | ||
| Input | $2/M | $0.07/M |
| Output | $10/M | $1.4/M |
| Cached input | $0.1/M | $0.069/M |
| Cache write | $0.188/M | — |
| Web search (per 1K calls) | $14/1K | — |
| Persian text | ||
| Tokens per 100 Persian words | 160 | 206 |
| Persian vs English tokens (same text) | ×1.4 | ×1.8 |
| Input cost of 1,000 Persian words | $0.0032 | $0.00014 |
| Features | ||
| Thinking (reasoning) | Always on | Always on |
| Max output tokens | 33K | 131K |
| Tool use | Yes | Yes |
| Structured output (JSON) | Yes | Yes |
| Prompt caching | Yes | Yes |
| Capabilities | ||
| Capabilities | Vision · Tools · Streaming · Long context | Tools · Streaming · Long context |
| Benchmarks | ||
| Coding Index | — | 75 |
| Intelligence Index | 59 | 45 |
| Agentic Index | — | 53 |
| GPQA DiamondGraduate-level scientific reasoning | — | 92% |
| HLEHumanity's Last Exam | — | 42% |
| AA-LCRLong-context reasoning | — | 80% |
| CritPtResearch-level physics reasoning | — | 19% |
| GDPvalEconomically valuable tasks | — | 58% |
| AA OmniscienceBroad factual knowledge accuracy | — | 34% |
| Non-HallucinationRate of avoiding fabricated answers | — | 70% |
| SciCodePython for scientific computing | — | 59% |
Gemini 4 Carbon and GLM 5.3: side-by-side summary
Gemini 4 Carbon, from Google, has a 1,048,576-token context window and costs $2/M for input and $10/M for output on Hormouz (M = one million tokens).
GLM 5.3, from Z.ai, has a 1,048,576-token context window and costs $0.07/M for input and $1.4/M for output on Hormouz (M = one million tokens).
Both are available through one OpenAI-compatible API and one Hormouz key, and you switch between them by changing the model id in your request.
Where they rank
The compared models' scores beside a few models of similar standing.
Intelligence Index
- Claude Fable 5.153.4
- GPT-6 Astra52.7
- GPT-6.1 Sol51.8
- GPT-6.1 Terra51.3
- GLM 5.344.8
Coding Index
- Claude Fable 5.181.6
- GPT-6 Astra76.9
- GLM 5.374.8
Agentic Index
- Claude Fable 5.157.9
- GLM 5.353.1
Data from Artificial Analysis; scores out of 100.
FAQ
- Which has the longer context window, Gemini 4 Carbon or GLM 5.3?
- Gemini 4 Carbon and GLM 5.3 have the same context window: 1,048,576 tokens.
- Which is cheaper, Gemini 4 Carbon or GLM 5.3?
- GLM 5.3 has the lower input price. Gemini 4 Carbon: $2/M input, $10/M output and GLM 5.3: $0.07/M input, $1.4/M output (per million tokens, on Hormouz).
- Which scores higher on intelligence, Gemini 4 Carbon or GLM 5.3?
- Gemini 4 Carbon scores higher on the Artificial Analysis Intelligence Index: 59 for Gemini 4 Carbon and 44.8 for GLM 5.3.
- Do Gemini 4 Carbon and GLM 5.3 reason before answering?
- Gemini 4 Carbon always reasons before answering; the reasoning effort setting controls how much. GLM 5.3 always reasons before answering; the reasoning effort setting controls how much.
- Can I use all 2 of these models through one API?
- Yes. All of them are available on Hormouz through a single OpenAI-compatible API and one API key. Switch between them by changing the model id.
Updated 2026-11-11