Skip to main content

Model profile

GLM-4.6

Z.AIEstablishedReleased Sep 1, 2025
Data verified
Overall Score
55.12Public #85 of 200Verified #50 of 99
Arena Elo
1425
Eligible category ranks
1of 8
Price (1M tokens)
Not listedAPI pricing
Speed
Not listed
Context
200K

Evidence coverage

14 of 323 tracked benchmarks are published. 3 are verified and 11 provisional. 6 of 8 categories are measured.

Updated July 23, 2026Methodology
Published / tracked
14 / 323
Verified
3
Provisional
11
Categories with evidence
6 / 8

Evidence by category

  • Agentic1 benchmark
    Reported
  • Coding2 benchmarks
    Mixed evidence
  • Reasoning2 benchmarks
    Reported
  • Knowledge6 benchmarks
    Reported
  • Math2 benchmarks
    Verified
  • Multilingual0 benchmarks
    Not measured
  • Multimodal0 benchmarks
    Not measured
  • Inst. Following1 benchmark
    Reported
Open WeightSelf-hostReasoning
Confidence:
Low
base

GLM-4.6 ranks #85 out of 200 models on the public leaderboard with an overall score of 55.12/100. It also ranks #50 out of 99 on the verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

GLM-4.6 is a open weight model with a 200K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

This profile currently has 14 of 323 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Coding (#53). This performance profile makes it particularly well-suited for software development and code generation tasks.

Peer position

Exact provisional scores and ranks for the closest listed peers. A score can appear before a model clears the evidence threshold for a rank, so equal scores can have different rank states.

Range 55.055.47

  1. DeepSeek V4 Pro (High)
    DeepSeek
    #8155.47
    DeepSeek V4 Pro (High) is #81 with a score of 55.47.
    Compare
  2. DeepSeek V3.2
    DeepSeek
    #8255.4
    DeepSeek V3.2 is #82 with a score of 55.4.
    Compare
  3. Gemini 3.1 Pro
    Google
    #8355.3
    Gemini 3.1 Pro is #83 with a score of 55.3.
    Compare
  4. GPT-5 (medium)
    OpenAI
    #8455.15
    GPT-5 (medium) is #84 with a score of 55.15.
    Compare
  5. GLM-4.6Current model
    Z.AI
    #8555.12
    GLM-4.6 is #85 with a score of 55.12.
  6. Step 3.5 Flash
    StepFun
    #8655.1
    Step 3.5 Flash is #86 with a score of 55.1.
    Compare
  7. Kimi K2.7 Code
    Moonshot AI
    #8755.0
    Kimi K2.7 Code is #87 with a score of 55.0.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Coding57%
    Eligible cohort rank #53 of 122Category score 51.9

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticWeight 22%1 benchmarkReportedScore pending
CodingRank #53 of 122Percentile 57thWeight 20%2 benchmarksMixed sources51.9
ReasoningWeight 17%2 benchmarksReportedScore pending
KnowledgeWeight 12%6 benchmarksReportedScore pending
MathRank Not rankedWeight 5%2 benchmarksVerified27.7
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalWeight 12%0 benchmarksNot measuredNot measured
Inst. FollowingWeight 5%1 benchmarkReportedScore pending

Chatbot Arena performance

Scroll horizontally to inspect confidence intervals and vote counts.

Chatbot Arena Elo, confidence interval, and vote count by evaluation view
ViewEloConfidence intervalVotes
Text Overall1425±3.935,613
Coding1459±7.37,476
Math1420±12.82,106
Instruction Following1415±6.49,997
Creative Writing1402±8.65,082
Multi-turn1421±8.25,664
Hard Prompts1442±5.019,072
Hard Prompts (English)1448±6.59,608
Longer Query1433±6.78,746

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic1 benchmark
τ²-bench resultsReported

τ²-Bench Tool-Agent-User Evaluation

76.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Coding2 benchmarks
Vibe Code BenchBenchmark exact

Vibe Code Bench v1.1

3.09%Display only
Source: Vals AI: Vibe Code Bench v1.1Provenance: Vals Vibe Code Bench v1.1 reports this exact row under zai/glm-4.6; BenchLM stores it on the local vibeCodeBench key.
AA-SciCodeReported

Artificial Analysis SciCode

33.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Reasoning2 benchmarks
AA-LCRReported

Artificial Analysis Long Context Reasoning

26.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

0.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge6 benchmarks
Artificial Analysis Intelligence IndexReported
23.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

63.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

5.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexReported

Artificial Analysis Omniscience Index

-31.6%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience AccuracyReported

Artificial Analysis Omniscience Accuracy

20.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

66.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Math2 benchmarks
FrontierMath v2 (Tiers 1-3)Benchmark exact

FrontierMath v2 Tiers 1-3

3.819%Weighted 30%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tiers 1-3 at 3.819% for zai-org/GLM-4.6. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
FrontierMath v2 (Tier 4)Benchmark exact

FrontierMath v2 Tier 4

2.128%Weighted 10%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tier 4 at 2.128% for zai-org/GLM-4.6. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
Inst. Following1 benchmark
AA-IFBenchReported

Artificial Analysis IFBench

36.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Frequently Asked Questions

How does GLM-4.6 perform overall in AI benchmarks?

GLM-4.6 has 14 published benchmark scores on BenchLM, but it does not yet have enough non-generated coverage to receive a global overall rank.

Is GLM-4.6 good for knowledge and understanding?

GLM-4.6 has visible benchmark coverage in knowledge and understanding, but BenchLM does not currently assign it a global category rank there.

Is GLM-4.6 good for coding and programming?

GLM-4.6 ranks #53 out of 122 models in coding and programming benchmarks with an average score of 51.9. There are stronger options in this category.

Is GLM-4.6 good for mathematics?

GLM-4.6 has visible benchmark coverage in mathematics, but BenchLM does not currently assign it a global category rank there.

Is GLM-4.6 good for reasoning and logic?

GLM-4.6 has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is GLM-4.6 good for agentic tool use and computer tasks?

GLM-4.6 has visible benchmark coverage in agentic tool use and computer tasks, but BenchLM does not currently assign it a global category rank there.

Is GLM-4.6 good for instruction following?

GLM-4.6 has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.

Is GLM-4.6 open source?

Yes, GLM-4.6 is an open weight model created by Z.AI, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

Does GLM-4.6 have full benchmark coverage on BenchLM?

Not yet. GLM-4.6 currently has 14 published benchmark scores out of the 323 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of GLM-4.6?

GLM-4.6 has a published context window of 200K, which determines how much text it can process in a single interaction.

Last updated: July 23, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Choose with this week’s evidence

Join 2,000+ readers for ranking moves, new releases, pricing changes, and the evidence behind them.

Free. One email per week.