Z.ai, released Aug 14, 2026

GLM-5.3price, context, benchmarks and release details

Provisional: not enough results to rank yet Open weights 93% confidence 93 percent, High confidence, 5 of 7 expected sources in
67.7
SI Score
#24 of 142 ranked models
Input, per 1M tokens
$1.40Z.ai API pricingOfficial Z.ai global Standard uncached token rate; coding-plan subscriptions, cache and Batch excluded Retrieved Oct 9, 2026 · factual citation
Open source ↗
Output, per 1M tokens
$4.40Z.ai API pricingOfficial Z.ai global Standard uncached token rate; coding-plan subscriptions, cache and Batch excluded Retrieved Oct 9, 2026 · factual citation
Open source ↗
Context window
1Mmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Max output
131Kmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Released
Aug 14, 2026models.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗

How the score breaks down

Coding (weight 40 percent) 63.2
Math (weight 15 percent) 74.8
Preference (weight 15 percent) 79.5
Reasoning (weight 30 percent) 84.4

Weights: reasoning 30%, math 15%, coding 40%, preference 15%. Results use fixed 0–100 scales before averaging, and thin evidence is pulled toward 50. Method si-v3-retained-evidence-2, computed Oct 9, 2026, 06:15 UTC.

Around it on the leaderboard

  1. 22 Gemini 3.1 Pro Preview 68.1
  2. 23 MiniMax-M3 68.0
  3. 24 GLM-5.3 67.7
  4. 25 Claude Sonnet 5 67.7
  5. 26 Claude Opus 4.8 67.5

Full leaderboard

Benchmark results

25 benchmarks, 25 results

Each row shows the best published result. Where a model was tested at several settings, such as reasoning effort, open the row to see each one. Hover or tap a value for its source.

GPQA Diamondreasoning 90.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 24, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 90.9
Humanity's Last Exam (with tools)reasoning 62.5%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] max effort; with tools Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 62.5
LiveBench Reasoning: connectionsreasoning 95.3%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 95.3
LiveBench Reasoning: consecutive eventsreasoning 72.3%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 72.3
LiveBench Reasoning: logic with navigationreasoning 70%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 70.0
LiveBench Reasoning: spatialreasoning 96%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 96.0
LiveBench Reasoning: theory of mindreasoning 88.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 88.5
LiveBench Reasoning: zebra puzzlesreasoning 88.8%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 88.8
FrontierMath Tier 4 (v2)math 29.3%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 25, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 29.3
FrontierMath Tiers 1–3 (v2)math 68.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 25, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 68.8
LiveBench Math: AMPS Hardmath 99%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 99.0
LiveBench Math: competition mathmath 96.1%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 96.1
LiveBench Math: integralsmath 69%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 69.0
LiveBench Math: olympiadmath 87.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 87.5
LiveBench Math: simplifymath 65.8%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 65.8
OTIS Mock AIME 2024–2025math 91.1%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 24, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 91.1
LiveBench Coding: code completioncoding 80.4%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 80.4
LiveBench Coding: code generationcoding 77.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 77.5
LiveBench Coding: JavaScriptcoding 72.7%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 72.7
LiveBench Coding: Pythoncoding 60%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 60.0
LiveBench Coding: TypeScriptcoding 50%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 50.0
Terminal-Bench 2.1coding 88.2%Official model cards via models.devLab-reported; metric pass@1; transcribed by MIT models.dev catalog; not independently evaluated [variant] max effort; Claude Code 2.1.207; 2.1 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 88.2
Terminal-Bench 3.0coding 28.3%Official model cards via models.devLab-reported; metric avg@3; transcribed by MIT models.dev catalog; not independently evaluated [variant] max effort; Claude Code 2.1.207; 3.0 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 28.3
Terminal-Bench 4.0coding 41.8%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Claude Code; maxPublished Sep 3, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
Open source ↗ 41.8
LMArena Textpreference 1471.2 eloLMArena / ArenaPublished source fact [variant] text / overallPublished Oct 8, 2026 Retrieved Oct 9, 2026 · CC-BY-4.0
Open source ↗ 79.5

Normalization uses a fixed 0–100 scale for each unit, independent of other models. Compare evaluation conditions before reading a small gap as decisive. “Lab-reported” marks the provider's own published figure.

Details and sources

Open weights
YesHugging Face HubPublic Hub repo with weight files; gating/repo upload date is not release date Retrieved Oct 9, 2026 · factual metadata; model-specific licenses
Open source ↗
License
otherHugging Face HubPublished source fact Retrieved Oct 9, 2026 · factual metadata; model-specific licenses
Open source ↗
Input modalities
textmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
First seen by SuperIndex
Oct 8, 2026
Coverage
75% of expected source weight

Reported (5)

  • Epoch AI BenchmarkingOct 8, 2026
  • Official model cards via models.devOct 8, 2026
  • LiveBenchOct 8, 2026
  • LMArena / ArenaOct 8, 2026
  • Terminal-BenchOct 8, 2026

Awaiting (2)

  • ARC Prize13% of weight
  • Humanity’s Last Exam13% of weight

Confidence rises as pending sources publish. Some sources never cover some models, so confidence reaches 100% at 80% of expected weight.

What changed

All releases · RSS feed

Alerts on this device

What to be alerted about
RSS feed