Zhipu AIGLM 4.7
GLM 4.7 is a Zhipu AI open-weights model, released 22 Dec 2025. Its best configuration (thinking reasoning effort) ranks #328 of 785 on the BenchLeader Index at 52.1 ±7.1, in the upper half. The ± is the point: 392 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (67) and lowest in reasoning (44). At $0.923 per million tokens blended it is mid-priced. Output speed of 38 tokens per second puts it in the slowest quarter, with a first answer in 54.6 s. It has been measured at 3 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Dec 2025.stale: no new result in six months
- Blended price
- $0.923/M
- $0.530 in · $2.10 out
- Output speed
- 38 tok/s
- measured by Artificial Analysis
- First answer
- 55 s
- first token 4.41 s
- Context
- 205k
- Full answer
- 68 s
- median, reasoning included
- Cost per run
- –
- Released
- 22 Dec 2025
- Overall index52
- Reasoning44
- Agents & tools47
- Knowledge45
- Instruction following67
- Long context60
- Composite54
Versions
Zhipu AI has shipped 8 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
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. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 48.9 | #412 | 38 tok/s | 1.59 s | $0.0008 | Agents & tools 60 · Composite 49 · Instruction following 55 · Knowledge 40 · Long context 45 · Reasoning 41 |
| thinkingbest | 52.1 | #328 | 38 tok/s | 55 s | $0.0008 | Agents & tools 47 · Composite 54 · Instruction following 67 · Knowledge 45 · Long context 60 · Reasoning 44 |
| not stated | 51.9 | #334 | 25 tok/s | 4.41 s | $0.0009 | Agents & tools 38 · Coding 54 · Human preference 63 · Knowledge 48 · Maths 54 · Reasoning 56 |
How its index has moved
8 Sept 2026 to 21 Oct 2026- thinking
- not stated
- no reasoning
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. 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 | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamondeffort not stated | – | – | 83.3%#103 | Epoch AI Benchmarking Hub | |
| SimpleBencheffort not stated | – | – | 47.7%#47 | SimpleBench | |
| Arena Hard Promptseffort not stated | – | – | 1462#90 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 66.4%#332 | 85.9%#117 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 6.4%#394 | 27.4%#175 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in indexeffort not stated | – | – | 80.0%#67 | Vals AI | |
| CritPt | 0.0%#315 | 1.7%#206 | – | Artificial Analysis | |
| Chess Puzzlesnot in indexeffort not stated | – | – | 21.0%#88 | Epoch AI Benchmarking Hub | |
| CL-benchnot in indexeffort not stated | – | – | 15.9%#16 | Epoch AI Benchmarking Hub | |
| CL-bench Lifenot in indexeffort not stated | – | – | 10.9%#13 | Epoch AI Benchmarking Hub |
Coding
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| SciCodeeffort not stated | – | 45.1%#139 | SciCode | |
| Arena Codingeffort not stated | – | 1484#102 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | 1435#79 | Arena (LMArena) | |
| LiveCodeBencheffort not stated | – | 82.2%#52 | Vals AI | |
| SWE-bench (Vals)not in indexeffort not stated | – | 69.4%#63 | Vals AI | |
| ALE-Benchnot in indexeffort not stated | – | 399.5#108 | Epoch AI Benchmarking Hub | |
| Terminal-Bench 2.1 (AA)not in index | 45.3%#127 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | – | 33.4%#57 | Terminal-Bench | |
| Terminal-Bench Hard | 30.3%#112 | 31.8%#101 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 94.2%#33 | 95.9%#15 | – | Artificial Analysis | |
| MCP Atlaseffort not stated | – | – | 58.1%#28 | Scale AI SEAL | |
| τ³-Banking (AA) | – | 12.2%#133 | – | Artificial Analysis | |
| Terminal-Bench 2.0 (Vals)not in indexeffort not stated | – | – | 38.2%#36 | Vals AI | |
| Vending-Bench 2not in indexeffort not stated | – | – | 2376.8#50 | Andon Labs | |
| BrowseComp-Plusnot in indexeffort not stated | – | – | 33.3%#5 | BrowseComp-Plus |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| OTIS Mock AIME | 83.3%#106 | Epoch AI Benchmarking Hub | |
| ProofBench | 6.0%#69 | Vals AI | |
| AIME (Vals) | 93.3%#15 | Vals AI | |
| MGSM | 88.2%#50 | Vals AI | |
| Arena Maths | 1428#108 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| SimpleQA Verifiedeffort not stated | – | – | 32.2%#64 | Epoch AI Benchmarking Hub | |
| AA-Omniscience | -47.3#366 | -36.4#307 | – | Artificial Analysis | |
| MMLU-Proeffort not stated | – | – | 82.7%#72 | Vals AI | |
| LegalBencheffort not stated | – | – | 83.4%#54 | Vals AI | |
| CorpFineffort not stated | – | – | 46.4%#106 | Vals AI | |
| TaxEvaleffort not stated | – | – | 68.8%#95 | Vals AI | |
| MedQAeffort not stated | – | – | 93.7%#21 | Vals AI | |
| CaseLaw v2not in indexeffort not stated | – | – | 54.9%#37 | Vals AI | |
| MedCodenot in indexeffort not stated | – | – | 32.8%#84 | Vals AI | |
| MedScribenot in indexeffort not stated | – | – | 68.6%#94 | Vals AI | |
| AA-Omniscience: accuracynot in index | 23.6%#254 | 29.3%#191 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 7.1%#491 | 7.0%#495 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| IFBench | 54.6%#151 | 67.9%#83 | – | Artificial Analysis | |
| Arena Instruction Followingnot in indexeffort not stated | – | – | 1426#102 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1441#92 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1405#101 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1457#79 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1451#88 | Arena (LMArena) |
Long context
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 40.7%#358 | 71.0%#206 | Artificial Analysis |
Composite
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 143.6#94 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 17.4#263 | 22.2#203 | – | 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 |
|---|---|---|---|---|---|---|
| Mancer | 26 tok/s | 1.44 s | $0.700 | $2.50 | 131k | fp4 |
| Z.ai | 24 tok/s | 8.88 s | $0.600 | $2.20 | 203k | fp4 |
| NovitaAI | 23 tok/s | 4.51 s | $0.540 | $1.98 | 205k | fp8 |
| DeepInfra | 22 tok/s | 1.54 s | $0.400 | $1.75 | 203k | fp4 |
| Venice | 22 tok/s | 1.41 s | $0.400 | $1.93 | 198k | fp4 |
Price history
The price above, per 1M tokens, every day it changed.
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 | $0.0007 | 1.0 min |
| Summarize a 30-page report | 12,000 / 600 | $0.0076 | $0.0038 | 1.2 min |
| Code edit | 6,000 / 1,500 | $0.0063 | $0.0044 | 1.6 min |
| Agentic coding session | 60,000 / 4,000 | $0.040 | $0.021 | 2.7 min |
| Structured extraction | 2,000 / 200 | $0.0015 | $0.0009 | 59.9 s |
More from Zhipu AI
See also
Questions people ask
- What are GLM 4.7's benchmark scores?
- As of 24 Oct 2026, GLM 4.7 (thinking effort) scores 52.1 ±7.1 on the BenchLeader Index, strongest in instruction following (67) and weakest in reasoning (44). That rests on 7 independent benchmark results, and every score on this page links to its publisher.
- How does GLM 4.7 rank?
- GLM 4.7 (thinking effort) is #328 of 785 ranked configurations on 24 Oct 2026. 392 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does GLM 4.7 cost per 1M tokens?
- As of 24 Oct 2026, GLM 4.7 costs $0.530 per 1M input tokens and $2.10 per 1M output tokens, $0.923 blended (three input tokens per output token). The list price last changed on 24 Oct 2026.
- When was GLM 4.7 released?
- Zhipu AI released GLM 4.7 on 22 Dec 2025. BenchLeader has results for it at 3 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “GLM 4.7: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/glm-4-7.