Zhipu AIGLM-4.5-Air
Best configuration ranks #373 of 610 on the BenchLeader Index at 46.0 ±5.1. Last measured 2 Sept 2026. Released 28 Jul 2025.
- Blended price
- $0.425/M
- $0.200 in · $1.10 out
- Output speed
- 54 tok/s
- First answer
- 40 s
- first token 1.30 s
- Context
- 131k
- Overall index46
- Reasoning43
- Coding46
- Agents & tools51
- Knowledge32
- Instruction following40
- Human preference55
- Long context49
- Composite43
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | default | Source |
|---|---|---|
| Humanity's Last Exam | 8.1%#29 | Scale AI / CAIS |
| LMArena Hard Prompts | 1390#158 | LMArena |
| GPQA Diamond (AA)not in index | 73.3%#258 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 7.0%#302 | Artificial Analysis |
| Kagi LLM Benchmark | 43.0%#98 | Kagi LLM Benchmark |
Coding
| Benchmark | default | Source |
|---|---|---|
| GSO-Bench | 2.9%#28 | GSO-Bench |
| LMArena Coding | 1427#155 | LMArena |
Agents & tools
| Benchmark | default | Source |
|---|---|---|
| Terminal-Bench Hard | 20.4%#157 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 46.5%#199 | Artificial Analysis |
Knowledge
| Benchmark | default | Source |
|---|---|---|
| AA-Omniscience | -61.5#398 | Artificial Analysis |
| MultiNRC | 10.4%#40 | Scale AI SEAL |
Instruction following
| Benchmark | default | Source |
|---|---|---|
| IFBench | 37.5%#277 | Artificial Analysis |
Human preference
| Benchmark | default | Source |
|---|---|---|
| LMArena Text | 1373#159 | LMArena |
Long context
| Benchmark | default | Source |
|---|---|---|
| AA-LCR | 46.7%#262 | Artificial Analysis |
Composite
| Benchmark | default | Source |
|---|---|---|
| AA Intelligence Index | 11.1#295 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| NovitaAI | 47 tok/s | 0.63 s | $0.130 | $0.850 | 131k | bf16 |
| Z.ai | 42 tok/s | 2.48 s | $0.200 | $1.10 | 131k | fp8 |
| SiliconFlow | 28 tok/s | 1.30 s | $0.140 | $0.860 | 131k | fp8 |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.030 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0004 | $0.0004 | 45.0 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0031 | $0.0015 | 50.6 s |
| Code edit | 6,000 / 1,500 | $0.0029 | $0.0021 | 1.1 min |
| Agentic coding session | 60,000 / 4,000 | $0.016 | $0.0088 | 1.9 min |
| Structured extraction | 2,000 / 200 | $0.0006 | $0.0004 | 43.2 s |
See also
Data as of 9 Sept 2026. Compare with another model.