GLM-5.3 Scores 60 on AA Intelligence Index, Tops Agentic Index

Zhipu's GLM-5.3 scored 60 on the Artificial Analysis Intelligence Index, up 7 points from GLM-5.2, tying Kimi K3 and just 1 point behind GPT-5.6, ranking 8/181; it also topped the AA Agentic Index with 59. It keeps the same 750B base as GLM-5.2, improved mainly through post-training and RL; the API is live, targeting complex coding, long-horizon agent tasks and defensive cybersecurity. If weights ship as planned, it would tie Kimi K3 as the leading open-weight model.

Confirmed

Intelligence Index score of 60, tying Kimi K3, up 7 from GLM-5.2, 1 point behind GPT-5.6 (@ArtificialAnlys, @Facelessjoe, @smlpth)
In Artificial Analysis' detailed review, GLM-5.3 (max) ranked 8/181, far above the tier median of 35; only 6 models score above 60; the evaluation generated 170M tokens; long context starts at $1.4 per million tokens
Strong agent gains with an ELO jump of 246; tops AA Agentic Index at 59, matching Fable/Opus 5, GPT 5.6 Sol and Kimi K3 at a fraction of their cost (@zainhas)
Terminal-Bench 3.0 score of 32.4%, roughly on par with GPT-5.6 Sol (34.6%) and Fable 5 (34.1%) (@zainhas); separate user testing found it close to Fable 5 (@scaling01)
Same 750B scale as GLM-5.2, improved via post-training and RL, with efficiency rivaling the near-3T-parameter Kimi K3; smallest model above 60 (@haider1, @卡尔的AI沃茨)
Zhipu launched the GLM-5.3 API for complex coding, long-horizon agents and defensive cybersecurity (@卡尔的AI沃茨)
@teortaxesTex relays ZixuanLi's report: the model extends its coding edge and shows notable gains in reasoning, general dialogue, and legal/finance domains, remaining among China's strongest
@zainhas notes the score looks almost unreal for a 700B base model, reflecting user surprise

Unconfirmed

Weight release remains a plan: multiple posters note that if shipped as planned, GLM-5.3 would tie Kimi K3 as the leading open-weight model by evals, but it had not landed as of posting

Why it matters

If weights ship as planned, GLM-5.3 and Kimi K3 would share the open-weight crown, reshaping the open-model landscape
Matching near-GPT-5.6-level performance without a larger base—via post-training alone—with a 246-point ELO jump and the top Agentic Index spot shows Chinese models rapidly closing in on frontier closed models on agentic tasks, corroborated by Terminal-Bench 3.0 results
As the smallest model in the 60+ tier, its price-performance and efficiency pressure same-tier competitors

2026-08-19 ~ 2026-08-20 · 12 related posts

Full story(15 episodes)→

Episode 1: Zhipu GLM-5.3 Spotted in Codebase, Post-Training Targets K3 and GPT-5.6(2026-08-03, 5 posts)
Episode 2: Zhipu AI Releases GLM-5.3: Same Base Model, Post-Training Drives a Big Leap(2026-08-14, 51 posts)
Episode 3: AI Weekly: Grok, Gemini Updates, and Anthropic's Secret Model(2026-08-14, 2 posts)
Episode 4: Alibaba, Zhipu and DeepSeek Unveil New Models on the Same Day(2026-08-15, 3 posts)
Episode 5: Zhipu Launches GLM-5.3 API with 50% Coding Boost(2026-08-19, 4 posts)
Episode 6: GLM-5.3 Scores 60 on AA Intelligence Index, Tops Agentic Index(2026-08-19, 12 posts)
Episode 7: Zhipu GLM-5.3 Tops DeepSWE at a Fraction of the Cost(2026-08-20, 4 posts)
Episode 8: Zhipu Open-Sources GLM-5.3-Flash: 320B MoE Multimodal at 1/10th the Cost(2026-08-25, 48 posts)
Episode 9: GLM-5.3 Flash Benchmarks Hit GPT-5.6 Territory at Fractional Cost(2026-08-27, 7 posts)
Episode 10: GLM-5.3-Flash Ranks Third Among Open Models at Half of Qwen's Cost(2026-08-27, 2 posts)
Episode 11: GLM 5.3 Flash Gets 90% Discount Through September(2026-08-27, 2 posts)
Episode 12: Zai Releases GLM-5.3 Weights with Tiered Access Citing Cybersecurity Capabilities(2026-08-27, 11 posts)
Episode 13: GLM-5.3 Flash Benchmarks: Near-Frontier Quality at a Fraction of the Cost(2026-08-27, 12 posts)
Episode 14: Z.ai Releases Open-Weight GLM-5.3 with Day-One Adoption Across Platforms(2026-08-29, 19 posts)
Episode 15: Zhipu's GLM-5.3 Launches Exclusively on Baseten with Big Coding Gains(2026-09-05, 2 posts)

Primary sources

GLM-5.3 ties Kimi K3 with score of 60, weights to be released — ArtificialAnlys · 2026-08-19 05:44 (UTC+08:00)
GLM-5.3 Benchmarks: Intel Score 60, 1M Context Window at $1.4 per Million Input — ArtificialAnlys · 2026-08-19 05:44 (UTC+08:00)
GLM-5.3 tops AA Agentic Index, matching frontier models at a fraction of the cost — zainhas · 2026-08-19 06:34 (UTC+08:00)

GLM-5.3 Scores 60 on AI Index, Touted as Strongest Chinese Model — teortaxesTex · 2026-08-19
[source] GLM-5.3 ties Kimi K3 with score of 60, weights to be released — ArtificialAnlys · 2026-08-19
[source] GLM-5.3 Benchmarks: Intel Score 60, 1M Context Window at $1.4 per Million Input — ArtificialAnlys · 2026-08-19
[source] GLM-5.3 tops AA Agentic Index, matching frontier models at a fraction of the cost — zainhas · 2026-08-19
GLM 5.3 scores 60 on AA overall intelligence index — zainhas · 2026-08-19
Zhipu GLM-5.3 scores 60, tying with Kimi K3 on Artificial Analysis — Facelessjoe · 2026-08-19
GLM 5.3 ranks among leaders in new benchmarks — markjeffrey · 2026-08-19
GLM-5.3 ties Kimi K3 with score 60, leading the open weights model rankings — smlpth · 2026-08-19
GLM-5.3 hits 60 on AA index, matching 3T param models with efficiency — haider1 · 2026-08-19
GLM-5.3 hands-on: 750B post-training lands top-10 on AA board, matches Kimi K3 in coding — 卡尔的AI沃茨 · 2026-08-20
GLM-5.3 crushes Terminal-Bench 3.0, nearly on par with Fable 5 — scaling01 · 2026-08-20
GLM 5.3 matches GPT-5.6 benchmark score at lower cost — zainhas · 2026-08-20