GLM 4.5
Z.ai · released Jul 25, 2025
GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications. It leverages a Mixture-of-Experts (MoE) architecture and supports a context length of up to 128k tokens. GLM-4.5 delivers significantly...
Specification
- Context window
- 131K
- Max output
- 98K
- Knowledge cutoff
- Dec 31, 2024
- Parameters
- 358.3B
- Licence
- mit
- Serving providers
- 1
- Moderated
- No
- Uptime
- 100.0%
Intelligence
19.7
59th percentile
Coding
—
Coding Index
Agentic
—
Agentic Index
Output speed
39 t/s
Median across providers
Latency
15.43s
Time to first token
Cost per task
$0.21
Estimated
Benchmarks
Where the score comes from
The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.
Evaluation scores
Percentage correct · higher is better
- MMLU-Pro83.5%
- GPQA Diamond78.2%
- LiveCodeBench73.8%
- AIME 202573.7%
- AA-LCR (long context)51.7%
- IFBench44.1%
- τ²-bench (Telecom)43.0%
- SciCode34.8%
- Terminal-Bench Hard22.0%
- Humanity's Last Exam13.0%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| MMLU-Pro | 83.5% |
| GPQA Diamond | 78.2% |
| LiveCodeBench | 73.8% |
| AIME 2025 | 73.7% |
| AA-LCR (long context) | 51.7% |
| IFBench | 44.1% |
| τ²-bench (Telecom) | 43.0% |
| SciCode | 34.8% |
| Terminal-Bench Hard | 22.0% |
| Humanity's Last Exam | 13.0% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- R1DeepSeek20.4
- North Mini Code (free)Cohere20.2
- GPT-5 NanoOpenAI20.1
- GLM 4.5Z.ai19.7
- Kimi K2 0711Moonshot AI19.7
- GPT-4.1OpenAI19.6
- o3 MiniOpenAI19.2
- o1-proOpenAI19.1
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| R1 | 20.4 |
| North Mini Code (free) | 20.2 |
| GPT-5 Nano | 20.1 |
| GLM 4.5 | 19.7 |
| Kimi K2 0711 | 19.7 |
| GPT-4.1 | 19.6 |
| o3 Mini | 19.2 |
| o1-pro | 19.1 |
Percentile among all indexed models
Pricing
What it costs to run
List prices per million tokens, plus what one representative task works out to.
List price
- Input / 1M tokens
- $0.6
- Output / 1M tokens
- $2.2
- Cached input / 1M
- $0.11
- Blended 3:1
- $1
One task, estimated
$0.21
- Input tokens
- 50,000
- Output tokens
- 80,000
- Profile
- Reasoning
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).
Serving providers
Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.
Arena
Head-to-head generation quality
Elo from pairwise judgements, broken out by the kind of thing the model was asked to build.
- Overall Elo
- 1,225
- Win rate
- 59.7%
- Strongest at
- 3D scenes
- Tournaments
- 1,878
Elo by category
Dot position on a 1,125–1,225 scale · Elo has no meaningful zero
- 1,225
- 1,192
- 1,186
- 1,186
- 1,179
- 1,143
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| 3D scenes | 1225 | 59.7% |
| Websites | 1192 | 53.8% |
| Data visualisation | 1186 | 53.3% |
| Game development | 1186 | 54.4% |
| UI components | 1179 | 55.0% |
| SVG | 1143 | 50.8% |