Zhipu AIGLM-4.5V
Best configuration ranks #449 of 610 on the BenchLeader Index at 42.9 ±6.3. Last measured 2 Sept 2026. Released 11 Aug 2025.
- Blended price
- $0.900/M
- $0.600 in · $1.80 out
- Output speed
- 37 tok/s
- First answer
- 2.71 s
- first token 2.56 s
- Context
- 64k
- Overall index43
- Reasoning54
- Coding54
- Agents & tools40
- Knowledge38
- Instruction following33
- Human preference53
- Multimodal37
- Long context25
- Composite37
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| thinking | 40.6 | #503 | 37 tok/s | 57 s | $0.0008 | Agents & tools 38 · Composite 39 · Instruction following 37 · Knowledge 42 · Long context 25 · Multimodal 34 |
| defaultbest | 42.9 | #449 | 37 tok/s | 2.71 s | $0.0008 | Agents & tools 40 · Coding 54 · Composite 37 · Human preference 53 · Instruction following 33 · Knowledge 38 · Long context 25 · Multimodal 37 · Reasoning 54 |
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 | thinking | default | Source |
|---|---|---|---|
| LMArena Hard Prompts | – | 1374#171 | LMArena |
| GPQA Diamond (AA)not in index | 68.4%#299 | 57.3%#373 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 6.3%#324 | 3.5%#508 | Artificial Analysis |
| Kagi LLM Benchmark | – | 59.8%#51 | Kagi LLM Benchmark |
Coding
| Benchmark | thinking | default | Source |
|---|---|---|---|
| LMArena Coding | – | 1404#169 | LMArena |
Agents & tools
| Benchmark | thinking | default | Source |
|---|---|---|---|
| Terminal-Bench Hard | 5.3%#263 | 6.8%#240 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 22.5%#297 | 19.6%#323 | Artificial Analysis |
Knowledge
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-Omniscience | -46.3#290 | -55.6#363 | Artificial Analysis |
Instruction following
| Benchmark | thinking | default | Source |
|---|---|---|---|
| IFBench | 34.2%#306 | 28.6%#352 | Artificial Analysis |
Human preference
| Benchmark | thinking | default | Source |
|---|---|---|---|
| LMArena Text | – | 1353#177 | LMArena |
Multimodal
| Benchmark | thinking | default | Source |
|---|---|---|---|
| LMArena Vision | – | 1155#96 | LMArena |
| MMMU-Pro | 50.5%#205 | 42.8%#225 | Artificial Analysis |
Long context
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-LCR | 0.0%#419 | 0.0%#419 | Artificial Analysis |
Composite
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA Intelligence Index | 7.6#387 | 6.7#428 | 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 | 38 tok/s | 2.00 s | $0.600 | $1.80 | 66k | fp8 |
| Z.ai | 37 tok/s | 3.11 s | $0.600 | $1.80 | 66k | 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 $0.110 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0008 | $0.0006 | 10.7 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0083 | $0.0039 | 18.8 s |
| Code edit | 6,000 / 1,500 | $0.0063 | $0.0041 | 42.8 s |
| Agentic coding session | 60,000 / 4,000 | $0.043 | $0.021 | 1.8 min |
| Structured extraction | 2,000 / 200 | $0.0016 | $0.0008 | 8.1 s |
See also
Data as of 9 Sept 2026. Compare these configurations.