Cheaper overall
Deepseek v4 Pro-0813
8% lower blended rate
Higher benchmark scores
Deepseek v4 Pro-0813
leads on 10 of 11 figures
Bigger context window
Deepseek v4 Pro-0813
1,048,576 vs 200,000 tokens
Cheaper cached input
Deepseek v4 Pro-0813
$0.044 vs $0.26, 83% less
Where each one wins
Deepseek v4 Pro-0813
- Cheaper input tokens $1.32 vs $1.40, 6% less
- Cheaper output tokens $3.96 vs $4.40, 10% less
- Cheaper cached input $0.044 vs $0.26, 83% less
- Cheaper on all four workloads
- Larger context window 1,048,576 tokens
- Longer maximum output 384,000 tokens
- Higher overall intelligence (Intelligence Index) 36 vs 26.1
- Higher coding ability (Coding Index) 68.8 vs 55.8
- Higher multi-step tool use (Agentic Index) 41.3 vs 23.9
- Ahead on 7 more benchmark figures
GLM 5.1
- Higher answer reliability (Omniscience Non-Hallucination) 70.1 vs 5.2
What four workloads cost
One run, and the same run a thousand times.
Deepseek v4 Pro-0813 is cheaper on all four workloads, by much the same margin each time — about 1.4 times, from $0.0033 against $0.0036 on a chat turn to $0.1346 against $0.2736 on a long document.
| Workload | Deepseek v4 Pro-0813 | GLM 5.1 | ×1,000 | Difference |
|---|---|---|---|---|
| Chat turn 1,000 in / 500 out | $0.0033 | $0.0036 | $3.30 / $3.60 | Deepseek v4 Pro-0813 is 8% cheaper |
| RAG answer 10,000 in / 800 out | $0.0164 | $0.0175 | $16.37 / $17.52 | Deepseek v4 Pro-0813 is 7% cheaper |
| Agent loop 32,000 in / 700 out × 20 calls | $0.1346 | $0.2736 | $134.64 / $273.60 | Deepseek v4 Pro-0813 is 51% cheaper |
| Long document 400,000 in / 3,000 out | GLM 5.1's context window holds 200,000 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 | Deepseek v4 Pro-0813 | GLM 5.1 | Difference |
|---|---|---|---|---|
| Any prompt length | Input | $1.32 | $1.40 | Deepseek v4 Pro-0813 is 6% cheaper |
| Any prompt length | Cached input | $0.044 | $0.26 | Deepseek v4 Pro-0813 is 83% cheaper |
| Any prompt length | Output | $3.96 | $4.40 | Deepseek v4 Pro-0813 is 10% cheaper |
Blended: Deepseek v4 Pro-0813 $1.98, GLM 5.1 $2.15 per 1M — one rate at 3:1 input to output, for comparing two models at a glance.
Deepseek v4 Pro-0813 priced by Perplexity, GLM 5.1 by Z.ai. Standard tier, pay-as-you-go.
Benchmarks
They share 11 figures: Deepseek v4 Pro-0813 leads on 10, GLM 5.1 on 1. The widest gap is 64.9 points, on answer reliability (Omniscience Non-Hallucination).
Overall intelligence
Intelligence Index — Overall intelligence across reasoning, knowledge & math evals
Deepseek v4 Pro-0813 ahead by 9.9
Coding ability
Coding Index — Coding ability across software-engineering evals
Deepseek v4 Pro-0813 ahead by 13
Multi-step tool use
Agentic Index — Tool use & multi-step agent task performance
Deepseek v4 Pro-0813 ahead by 17.4
Graduate-level science
GPQA Diamond — Graduate-level questions in biology, chemistry & physics
Deepseek v4 Pro-0813 ahead by 6
Expert-exam performance
Humanity's Last Exam — Expert-level questions across many academic domains
Deepseek v4 Pro-0813 ahead by 10.9
Long-context reasoning
AA-LCR — Long-context reasoning across large inputs
Deepseek v4 Pro-0813 ahead by 6.6
Scientific coding
SciCode — Research-level scientific coding tasks
Deepseek v4 Pro-0813 ahead by 6.2
Physics research reasoning
CritPt — Unpublished physics research reasoning problems
Deepseek v4 Pro-0813 ahead by 13.4
Real-world task performance
GDPval — Economically valuable, real-world knowledge work
Deepseek v4 Pro-0813 ahead by 17.1
Factual accuracy
Omniscience Accuracy — Breadth of factual knowledge across domains
Deepseek v4 Pro-0813 ahead by 23.9
Answer reliability
Omniscience Non-Hallucination — How reliably the model avoids fabricated answers
GLM 5.1 ahead by 64.9
Each figure is that model's best published run, with the effort level named beside it — where these scores come from.
Specifications
Deepseek v4 Pro-0813 holds 848,576 more input tokens in one request. Only Deepseek v4 Pro-0813 lists code execution.
| Specification | Deepseek v4 Pro-0813 | GLM 5.1 |
|---|---|---|
| Context window | 1,048,576 tokens | 200,000 tokens |
| Max output | 384,000 tokens | 128,000 tokens |
| Takes in | Text | Text |
| Puts out | Text | Text |
| Released | 12 Aug 2026 | 7 Apr 2026 |
| Status | Active | Active |
| Sold by | Perplexity | Z.ai |
| Tool use | Yes | Yes |
| Structured outputs | Yes | Yes |
| Web search | Not listed | Not listed |
| Prompt caching | Yes | Yes |
| Code execution | Yes | Not listed |
| Computer use | Not listed | Not listed |
| Open weights | Yes | Yes |
| Reasoning | Yes | Yes |
FAQs
- Is Deepseek v4 Pro-0813 cheaper than GLM 5.1?
- Yes. A 1,000-token prompt with a 500-token reply costs $0.0033 on Deepseek v4 Pro-0813 and $0.0036 on GLM 5.1, and the same model is cheaper on every workload on this page.
- How much do Deepseek v4 Pro-0813 and GLM 5.1 cost per 1M tokens?
- Deepseek v4 Pro-0813 costs $1.32 for input and $3.96 for output. GLM 5.1 costs $1.40 and $4.40. Standard pay-as-you-go rates, as of 17 Oct 2026.
- Which is better, Deepseek v4 Pro-0813 or GLM 5.1?
- On published benchmarks, Deepseek v4 Pro-0813. It leads on 10 of the 11 figures both models report, including overall intelligence (Intelligence Index), where it scores 36 against 26.1.
- Does Deepseek v4 Pro-0813 or GLM 5.1 have a bigger context window?
- Deepseek v4 Pro-0813, at 1,048,576 tokens against 200,000. That is 848,576 more input tokens in a single request.