Zhipu AIGLM-4.7
Best configuration ranks #187 of 610 on the BenchLeader Index at 54.9 ±4.0. Last measured 8 Sept 2026. Released 22 Dec 2025.
- Blended price
- $1.00/M
- $0.600 in · $2.20 out
- Output speed
- 94 tok/s
- First answer
- 23 s
- first token 1.71 s
- Context
- 205k
- Overall index55
- Reasoning56
- Coding53
- Agents & tools42
- Maths52
- Knowledge48
- Instruction following66
- Human preference63
- Long context62
- Composite57
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 |
|---|---|---|---|---|---|---|
| no reasoning | 50.4 | #286 | 96 tok/s | 1.40 s | $0.0009 | Agents & tools 60 · Composite 51 · Instruction following 55 · Knowledge 42 · Long context 46 |
| defaultbest | 54.9 | #187 | 94 tok/s | 23 s | $0.0009 | Agents & tools 42 · Coding 53 · Composite 57 · Human preference 63 · Instruction following 66 · Knowledge 48 · Long context 62 · Maths 52 · Reasoning 56 |
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 | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| GPQA Diamond | – | 83.3%#91 | Epoch AI Benchmarking Hub | |
| SimpleBench | – | 47.7%#47 | SimpleBench | |
| LMArena Hard Prompts | – | 1462#75 | LMArena | |
| GPQA Diamond (AA)not in index | 66.4%#317 | 85.9%#116 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 6.4%#321 | 27.4%#127 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | 80.0%#67 | Vals AI |
Coding
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| SciCode | – | 45.1%#86 | SciCode | |
| LMArena Coding | – | 1485#85 | LMArena | |
| LMArena WebDev | – | 1435#61 | LMArena | |
| LiveCodeBench | – | 82.2%#52 | Vals AI | |
| IOI | – | 7.6%#37 | Vals AI | |
| SWE-bench (Vals)not in index | – | 69.4%#63 | Vals AI |
Agents & tools
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| Terminal-Bench | – | 33.4%#43 | Terminal-Bench | |
| APEX-Agents | – | 8.7%#54 | Mercor | |
| Terminal-Bench Hard | 30.3%#110 | 31.8%#99 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 94.2%#34 | 95.9%#16 | Artificial Analysis | |
| MCP Atlas | – | 58.1%#25 | Scale AI SEAL |
Maths
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| OTIS Mock AIME | – | 83.3%#91 | Epoch AI Benchmarking Hub | |
| ProofBench | – | 6.0%#53 | Vals AI | |
| AIME (Vals) | – | 93.3%#15 | Vals AI | |
| MGSM | – | 88.2%#51 | Vals AI |
Knowledge
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| SimpleQA Verified | – | 32.2%#54 | Epoch AI Benchmarking Hub | |
| AA-Omniscience | -47.3#295 | -36.4#240 | Artificial Analysis | |
| MMLU-Pro | – | 82.7%#73 | Vals AI | |
| LegalBench | – | 83.4%#50 | Vals AI | |
| CorpFin | – | 46.4%#107 | Vals AI | |
| TaxEval | – | 68.8%#96 | Vals AI | |
| MedQA | – | 93.7%#21 | Vals AI |
Instruction following
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| IFBench | 54.6%#148 | 67.9%#82 | Artificial Analysis |
Human preference
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| LMArena Text | – | 1442#78 | LMArena |
Long context
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 40.7%#288 | 71.0%#152 | Artificial Analysis |
Composite
| Benchmark | no reasoning | default | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 143.5#80 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | 17.4#200 | 22.2#151 | 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 |
|---|---|---|---|---|---|---|
| Google Vertex | 87 tok/s | 0.55 s | $0.600 | $2.20 | 200k | – |
| AtlasCloud | 56 tok/s | 1.06 s | $0.520 | $1.85 | 203k | fp8 |
| Z.ai | 28 tok/s | 15 s | $0.600 | $2.20 | 203k | fp4 |
| NovitaAI | 20 tok/s | 3.70 s | $0.540 | $1.98 | 205k | fp8 |
| DeepInfra | 15 tok/s | 1.27 s | $0.400 | $1.75 | 203k | fp4 |
| Venice | 13 tok/s | 2.50 s | $0.550 | $2.65 | 198k | fp4 |
| Mancer | 13 tok/s | 1.71 s | $0.700 | $2.50 | 131k | fp4 |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
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 | 25.9 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0085 | $0.0041 | 29.1 s |
| Code edit | 6,000 / 1,500 | $0.0069 | $0.0047 | 38.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.045 | $0.023 | 1.1 min |
| Structured extraction | 2,000 / 200 | $0.0016 | $0.0009 | 24.8 s |
See also
Data as of 9 Sept 2026. Compare these configurations.