Cheaper overall
GLM 4.6V
55% lower blended rate
Higher benchmark scores
GLM 4.6
leads on 9 of 10 figures
Bigger context window
GLM 4.6
204,800 vs 128,000 tokens
Cheaper cached input
GLM 4.6V
$0.05 vs $0.11, 55% less
Where each one wins
GLM 4.6V
- Cheaper input tokens $0.30 vs $0.60, 50% less
- Cheaper output tokens $0.90 vs $2.20, 59% less
- Cheaper cached input $0.05 vs $0.11, 55% less
- Cheaper on all four workloads
- Longer maximum output 32,768 tokens
- Higher answer reliability (Omniscience Non-Hallucination) 48.5 vs 32.4
GLM 4.6
- Larger context window 204,800 tokens
- Higher overall intelligence (Intelligence Index) 18.5 vs 11.2
- Higher graduate-level science (GPQA Diamond) 78 vs 71.9
- Higher expert-exam performance (Humanity's Last Exam) 14.5 vs 9.6
- Ahead on 6 more benchmark figures
What four workloads cost
One run, and the same run a thousand times.
GLM 4.6V is cheaper on all four workloads, by much the same margin each time โ about 2.2 times, from $0.0008 against $0.0017 on a chat turn to $0.0546 against $0.1208 on a long document.
| Workload | GLM 4.6V | GLM 4.6 | ร1,000 | Difference |
|---|---|---|---|---|
| Chat turn 1,000 in / 500 out | $0.0008 | $0.0017 | $0.7500 / $1.70 | GLM 4.6V is 56% cheaper |
| RAG answer 10,000 in / 800 out | $0.0037 | $0.0078 | $3.72 / $7.76 | GLM 4.6V is 52% cheaper |
| Agent loop 32,000 in / 700 out ร 20 calls | $0.0546 | $0.1208 | $54.60 / $120.80 | GLM 4.6V is 55% cheaper |
| Long document 400,000 in / 3,000 out | GLM 4.6V's context window holds 128,000 tokens, so a 400,000-token prompt does not fit. | |||
Price per 1M tokens
Neither model changes its rate with prompt length, so these rates apply to every request.
| Context band | Price | GLM 4.6V | GLM 4.6 | Difference |
|---|---|---|---|---|
| Any prompt length | Input | $0.30 | $0.60 | GLM 4.6V is 50% cheaper |
| Any prompt length | Cached input | $0.05 | $0.11 | GLM 4.6V is 55% cheaper |
| Any prompt length | Output | $0.90 | $2.20 | GLM 4.6V is 59% cheaper |
Blended: GLM 4.6V $0.45, GLM 4.6 $1.00 per 1M โ one rate at 3:1 input to output, for comparing two models at a glance.
Both priced by Z.ai. Standard tier, pay-as-you-go.
Benchmarks
They share 10 figures: GLM 4.6V leads on 1, GLM 4.6 on 9. The widest gap is 45.3 points, on support-agent performance (ฯยฒ-Bench Telecom).
Overall intelligence
Intelligence Index โ Overall intelligence across reasoning, knowledge & math evals
GLM 4.6 ahead by 7.3
Graduate-level science
GPQA Diamond โ Graduate-level questions in biology, chemistry & physics
GLM 4.6 ahead by 6.1
Expert-exam performance
Humanity's Last Exam โ Expert-level questions across many academic domains
GLM 4.6 ahead by 4.9
Instruction following
IFBench โ Precise following of detailed instructions
GLM 4.6 ahead by 13.3
Support-agent performance
ฯยฒ-Bench Telecom โ Tool-using agent tasks in a telecom support setting
GLM 4.6 ahead by 45.3
Long-context reasoning
AA-LCR โ Long-context reasoning across large inputs
GLM 4.6 ahead by 5.3
Command-line work
Terminal-Bench Hard โ Complex command-line and terminal workflows
GLM 4.6 ahead by 14.4
Physics research reasoning
CritPt โ Unpublished physics research reasoning problems
GLM 4.6 ahead by 1.1
Factual accuracy
Omniscience Accuracy โ Breadth of factual knowledge across domains
GLM 4.6 ahead by 9.5
Answer reliability
Omniscience Non-Hallucination โ How reliably the model avoids fabricated answers
GLM 4.6V ahead by 16.1
Each figure is that model's best published run, with the effort level named beside it โ where these scores come from.
Specifications
GLM 4.6 holds 76,800 more input tokens in one request. GLM 4.6V also takes file, image and video. Only GLM 4.6 lists structured outputs and prompt caching.
| Specification | GLM 4.6V | GLM 4.6 |
|---|---|---|
| Context window | 128,000 tokens | 204,800 tokens |
| Max output | 32,768 tokens | 16,384 tokens |
| Takes in | Text, image, video, file | Text |
| Puts out | Text | Text |
| Knowledge cutoff | โ | 31 Mar 2025 |
| Released | 8 Dec 2025 | 30 Sep 2025 |
| Status | Active | Active |
| Sold by | Z.ai | Z.ai |
| Tool use | Yes | Yes |
| Structured outputs | Not listed | Yes |
| Web search | Not listed | Not listed |
| Prompt caching | Not listed | Yes |
| Code execution | Not listed | Not listed |
| Computer use | Not listed | Not listed |
| Open weights | Yes | Yes |
| Reasoning | Yes | Yes |
FAQs
- Is GLM 4.6V cheaper than GLM 4.6?
- Yes. A 1,000-token prompt with a 500-token reply costs $0.0008 on GLM 4.6V and $0.0017 on GLM 4.6, and the same model is cheaper on every workload on this page.
- How much do GLM 4.6V and GLM 4.6 cost per 1M tokens?
- GLM 4.6V costs $0.30 for input and $0.90 for output. GLM 4.6 costs $0.60 and $2.20. Standard pay-as-you-go rates, as of 8 Dec 2025.
- Which is better, GLM 4.6V or GLM 4.6?
- On published benchmarks, GLM 4.6. It leads on 9 of the 10 figures both models report, including overall intelligence (Intelligence Index), where it scores 18.5 against 11.2.
- Does GLM 4.6V or GLM 4.6 have a bigger context window?
- GLM 4.6, at 204,800 tokens against 128,000. That is 76,800 more input tokens in a single request.