Zhipu AIGLM-4.5
Best configuration ranks #289 of 610 on the BenchLeader Index at 50.2 ±3.2. Last measured 2 Sept 2026. Released 28 Jul 2025.
- Blended price
- $1.00/M
- $0.600 in · $2.20 out
- Output speed
- 17 tok/s
- First answer
- 9.78 s
- Context
- 131k
- Overall index50
- Reasoning48
- Coding48
- Agents & tools53
- Maths51
- Knowledge50
- Instruction following46
- Human preference60
- Long context52
- Composite45
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 | – | – | 17 tok/s | 9.78 s | $0.0009 | Coding 44 |
| defaultbest | 50.2 | #289 | 17 tok/s | 9.78 s | $0.0009 | Agents & tools 53 · Coding 48 · Composite 45 · Human preference 60 · Instruction following 46 · Knowledge 50 · Long context 52 · Maths 51 · Reasoning 48 |
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 |
|---|---|---|---|
| Humanity's Last Exam | – | 8.3%#28 | Scale AI / CAIS |
| LMArena Hard Prompts | – | 1433#114 | LMArena |
| GPQA Diamond (AA)not in index | – | 78.2%#207 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | – | 13.0%#213 | Artificial Analysis |
| GPQA Diamond (Vals)not in index | – | 72.2%#85 | Vals AI |
| Kagi LLM Benchmark | – | 57.9%#55 | Kagi LLM Benchmark |
Coding
| Benchmark | thinking | default | Source |
|---|---|---|---|
| WeirdML | 40.6%#100 | – | WeirdML |
| LMArena Coding | – | 1455#122 | LMArena |
| LiveCodeBench | – | 67.5%#93 | Vals AI |
| IOI | – | 2.9%#52 | Vals AI |
| SWE-bench Verified (bash only) | – | 54.2%#30 | SWE-bench |
| SWE-bench Verified (any scaffold)not in index | – | 54.2%#38 | SWE-bench |
Agents & tools
| Benchmark | thinking | default | Source |
|---|---|---|---|
| Terminal-Bench Hard | – | 22.0%#149 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | – | 43.0%#211 | Artificial Analysis |
Maths
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AIME (Vals) | – | 86.7%#31 | Vals AI |
| MGSM | – | 90.8%#35 | Vals AI |
| MathArena Apex | – | 1.0%#35 | MathArena |
Knowledge
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-Omniscience | – | -27.4#208 | Artificial Analysis |
| MMLU-Pro | – | 81.2%#80 | Vals AI |
| LegalBench | – | 75.6%#104 | Vals AI |
| CorpFin | – | 61.0%#61 | Vals AI |
| TaxEval | – | 72.4%#62 | Vals AI |
| MedQA | – | 90.0%#49 | Vals AI |
| MultiNRC | – | 17.4%#37 | Scale AI SEAL |
Instruction following
| Benchmark | thinking | default | Source |
|---|---|---|---|
| IFBench | – | 44.1%#213 | Artificial Analysis |
Human preference
| Benchmark | thinking | default | Source |
|---|---|---|---|
| LMArena Text | – | 1411#120 | LMArena |
Long context
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-LCR | – | 52.7%#247 | Artificial Analysis |
Composite
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA Intelligence Index | – | 12.8#257 | 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 |
|---|---|---|---|---|---|---|
| Z.ai | 17 tok/s | 9.78 s | $0.600 | $2.20 | 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 $0.110 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0009 | $0.0008 | 27.4 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0085 | $0.0041 | 45.1 s |
| Code edit | 6,000 / 1,500 | $0.0069 | $0.0047 | 1.6 min |
| Agentic coding session | 60,000 / 4,000 | $0.045 | $0.023 | 4.1 min |
| Structured extraction | 2,000 / 200 | $0.0016 | $0.0009 | 21.5 s |
See also
Data as of 9 Sept 2026. Compare these configurations.