GLM 5
Z.ai · released Feb 11, 2026
GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows. Built for expert developers, it delivers production-grade performance on large-scale programming tasks, rivaling leading...
Specification
- Context window
- 205K
- Max output
- 131K
- Knowledge cutoff
- Not stated
- Parameters
- 753.9B
- Licence
- mit
- Serving providers
- 8
- Moderated
- No
- Uptime
- 100.0%
Intelligence
27.9
80th percentile
Coding
—
Coding Index
Agentic
—
Agentic Index
Output speed
113 t/s
Median across providers
Latency
1.05s
Time to first token
Cost per task
$0.31
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
- τ²-bench (Telecom)98.2%
- GPQA Diamond82.0%
- AA-LCR (long context)75.7%
- IFBench72.3%
- Terminal-Bench Hard43.2%
- Humanity's Last Exam29.3%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| τ²-bench (Telecom) | 98.2% |
| GPQA Diamond | 82.0% |
| AA-LCR (long context) | 75.7% |
| IFBench | 72.3% |
| Terminal-Bench Hard | 43.2% |
| Humanity's Last Exam | 29.3% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- MiniMax M3MiniMax29.2
- GPT-5.2-CodexOpenAI28.5
- Solar Pro 4Upstage28.2
- GLM 5Z.ai27.9
- Kimi K2.6Moonshot AI27.0
- Qwen3.6 PlusQwen27.0
- Ling-2.6-1TInclusionAI26.6
- GLM 5 TurboZ.ai26.6
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| MiniMax M3 | 29.2 |
| GPT-5.2-Codex | 28.5 |
| Solar Pro 4 | 28.2 |
| GLM 5 | 27.9 |
| Kimi K2.6 | 27.0 |
| Qwen3.6 Plus | 27.0 |
| Ling-2.6-1T | 26.6 |
| GLM 5 Turbo | 26.6 |
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
- $1
- Output / 1M tokens
- $3.2
- Cached input / 1M
- $0.2
- Blended 3:1
- $1.55
One task, estimated
$0.31
- 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
- 1,249
- Win rate
- 55.0%
- Strongest at
- Websites
- Tournaments
- 24,173
Elo by category
Dot position on a 1,100–1,250 scale · Elo has no meaningful zero
- 1,249
- 1,237
- 1,232
- 1,229
- 1,229
- 1,166
- 1,133
- 1,123
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| Websites | 1249 | 55.0% |
| Game development | 1237 | 57.4% |
| UI components | 1232 | 53.6% |
| Data visualisation | 1229 | 53.0% |
| 3D scenes | 1229 | 56.3% |
| SVG | 1166 | 54.4% |
| Mobile apps | 1133 | 51.6% |
| Full-stack apps | 1123 | 51.7% |