GLM 4.6
z-ai/glm-4.6- Z.ai
- Family · GLM 4
- Released 30 Sep 2025
- Open weights
- Text → Text
- Tools · JSON
Spec & list price
Context200K
Max output128K
Input / 1M$0.50
Output / 1M$2.004× input
Scores · standing
Intelligence18.5#152/489
Coding45.8#76/196
Usage
Tokens / day1.1B▼ 39% 7d
Usage rank#189
Requests / day63.2K
Tool calls / req0.16
Per request · 30 days
Avg input11.5K0.8× median
Avg output5470.8× median
Cache hit63%about median
Reasoning share40%0.7× median
Tool-call failures0.4%0.2× median
Consumption
Peak 3.3B 28 SepLow 816.5M 31 OctRank today #189
News & trends
3 mentions this week ▲ 200%GLM 4 family · videos, posts, articles · 30 days
Trending · 14 days · GLM 4 family
- DeepSeek 4.1 is the second model to be ranked in Bad. Where would you rank GLM 4.7?21 upvotes
- Stuck between GLM 4.7, 4.6, and Kimi K2.5. Also, which provider should I use between OpenRouter and DeepInfra?49 upvotes
News · latest from Z.ai
- Hermes and REA and GLM 5.3 = Reverse engineering TikTok app for less than $20scraping.club · 29 Oct
- GLM-5.3 Sometimes Considers How Train-Time Actions Shape Future Behaviorlesswrong.com · 27 Oct
- Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)blog.cloudflare.com · 26 Oct
- ReplaySSM: Rethinking State Storage for Faster Hybrid-Model DecodingNVIDIA · 12 Oct
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 5 providers, latest day
Output speed#300/33327 tok/s2.5× slower than median
Fastest provider31 tok/sbest of 5 providers
First token#2702.32 s2.3× longer than median
#16 cheapest of 21 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases22.2
Technical details
- Tokenizer
- Other
- Knowledge cutoff
- 2025-03-31
- License
- MIT
- Parameters
- 357B
- Moderation
- none
- Downloads · likes
- 26.2K · 1.2K
Request parameters 17
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- include_reasoning
- temperature
- top_p
- top_k
- min_p
+7 more
- max_tokens
- stop
- seed
- frequency_penalty
- logit_bias
- presence_penalty
- repetition_penalty
Interesting facts
- 2nd of 3 GLM 4 models by tokens last month.
- Weekdays average 2.5B tokens a day against 2.5B on weekends.
When to use
Best for
- General chat with tool calling
- Knowledge-heavy question answering
- Math-style reasoning
- Self-hosted assistants
Think twice for
- Image input
- Very long documents
- Newer agentic coding work
Research papers
- A2M: Trace-Optimized Agent Hijacking in the MCP Ecosystem22 Sep
- The Nuclear Decision-Making Benchmark: Evaluating Frontier LLMs on Nuclear Tendencies29 Jun
- OralMLLM-Bench: Evaluating Cognitive Capabilities of Multimodal Large Language Models in Dental Practice2 May
- SkillX: Automatically Constructing Skill Knowledge Bases for Agents6 Apr
- EnterpriseBench Corecraft: Training Generalizable Agents on High-Fidelity RL Environments18 Feb
- EGSS: Entropy-guided Stepwise Scaling for Reliable Software Engineering5 Feb
- daVinci-Agency: Unlocking Long-Horizon Agency Data-Efficiently2 Feb
- ROMA: Recursive Open Meta-Agent Framework for Long-Horizon Multi-Agent Systems2 Feb
Same family
GLM 4- 22.2
- 12.8
Alternatives
other vendors · similar capability- 21.8
- 21.5
- 22.2
- S19.5
Using GLM 4.6 at volume?
Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.