- π Z.ai's February 2026 flagship for agentic engineering and reasoning.
- π 744B-parameter MoE, 40B active, scaled up from GLM-4.5.
- π§ Trained on 28.5T tokens with new asynchronous RL infrastructure.
- π§ Adds DeepSeek Sparse Attention to cut training and inference cost.
- π Large context window (catalog: 198K tokens), FP8 weights available.
- π― Capabilities: reasoning, code-optimization, function calling, web search.
- π Released open-weight under the permissive MIT license.
- π¬ Vendor reports gains over GLM-4.7 across reasoning, coding, agentic tasks.
Z.ai, formally Knowledge Atlas Technology Joint Stock Co., Ltd., is a Chinese technology company specializing in artificial intelligence. Previously known internationally as Zhipu AI, the company rebranded to Z.ai in 2025. Its core focus is the GLM family of large languageβ¦
Explore 17 more models by Z.ai βGLM 5 is the February 2026 flagship from Z.ai (formerly Zhipu AI), positioned for complex systems engineering and long-horizon agentic work. Architecturally it is a Mixture-of-Experts model with 744 billion total parameters and roughly 40 billion active per token, scaled up from GLM-4.5's 355B (32B active), with pre-training data expanded to 28.5 trillion tokens. It is distributed open-weight under the MIT license in both full-precision and FP8 formats.
The two headline changes over earlier generations are efficiency-focused. GLM 5 adopts DeepSeek Sparse Attention (DSA), which the technical report describes as dynamically allocating attention by token importance to lower compute without compromising long-context understanding β an advance over the standard MoE used in GLM-4.5. Post-training uses a new asynchronous reinforcement-learning infrastructure built on the "slime" framework that decouples generation from training to improve GPU utilization.
Relative to its same-family predecessor GLM 4.7, Z.ai reports significant improvements across academic benchmarks in reasoning, coding, and agentic tasks.
GLM 5 was followed by refreshed siblings GLM 5.1 and GLM 5.2, the latter extending to a roughly 1M-token context with the IndexShare architecture.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usageΒ Β·Β GLM 5Β Β·Β Organic providers
Top providersΒ Β·Β GLM 5Β Β·Β Organic
- Open Forge99.0%
- Venice.ai Proxy0.8%
- β² Apex Ant0.1%
- Best0.0%
- Relay Works0.0%
- Night Harbor AI0.0%
- Others0.0%
| Provider | Health | Repβ | Sales | Tokens | Routing | Input $/M | Cached $/M | Output $/M | Save | 1st token | tok/s |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Super Seeder 0xd19fβ¦41f3 | healthy | 90.57 | $0.01 | 17.30k | #3 | $0.9427 | $0.1885 | $3.0166 | β6% | β | β |
| β² Apex Ant 0x73b4β¦e736 | healthy | 88.20 | $0.26 | 402.85k | #1 | $0.50 | $0.10 | $1.60 | β50% | β | β |
| Venice.ai Proxy 0x1f22β¦18c9 | healthy | 75.07 | $2.97 | 2.37M | #2 | $0.55 | $0.11 | $1.76 | β45% | β | β |
| Relay Works 0x3040β¦7139 | β | 58.03 | $0.06 | 91.29k | gated | $0.39 | $0.078 | $1.248 | β61% | β | β |
| HallPass 0xcac8β¦71b4 | β | 20.00 | β | β | gated | $0.018 | $0.003 | $0.06 | β98% | β | β |
| klinn 0x0ee1β¦6c41 | β | 12.34 | $0.02 | 2.75M | gated | $0.0057 | $0.0057 | $0.0184 | β99% | β | β |
| dealer 0xc46eβ¦8ce6 | β | 2.11 | $0.00 | 1.30k | gated | $0.95 | $0.19 | $3.04 | β5% | β | β |
| surplus-hive 0x6145β¦5c58 | β | 0.00 | β | β | gated | $1.5302 | $1.5302 | $4.7711 | β | β | β |
| uomi.ai 0x87dfβ¦48e3 | β | 0.00 | $0.00 | 195.00 | gated | $0.938 | $0.938not advertised | $2.889 | β6% | β | β |
| Fire Ant π₯π 0xbe05β¦bc5d | β | 0.00 | $0.03 | 80.36k | gated | $0.79 | $0.15 | $2.55 | β21% | β | β |
| AntFeed 0xddb6β¦1442 | β | 0.00 | β | β | gated | $0.66 | $0.66not advertised | $2.112 | β34% | β | β |
| Open Bird 0xc0f1β¦8183 | β | 0.00 | $0.00 | 440.00 | gated | $0.55 | $0.11 | $1.76 | β45% | β | β |
| antseed-neon-puma-944e 0x6650β¦944e | β | 0.00 | $0.01 | 22.38k | gated | $0.341 | $0.20 | $1.0912 | β66% | β | β |
| Apex TEE Test 0xe672β¦7955 | β | 0.00 | β | β | gated | $0.1482 | $0.0495 | $0.4743 | β85% | β | β |
| Best 0x472fβ¦69fd | healthy | 0.00 | $0.09 | 343.06k | gated | $0.1424 | $0.121 | $0.4557 | β86% | 0.2 s | 210 |
| Leftermute 0x388bβ¦5389 | β | 0.00 | β | β | gated | $0.10 | $0.10not advertised | $0.3772 | β90% | β | β |
| Night Harbor AI 0x305eβ¦fcf6 | β | 0.00 | $0.05 | 464.96k | gated | $0.08 | $0.08not advertised | $0.32 | β92% | β | β |
| Joi Gateway 0x8a90β¦7007 | β | 0.00 | β | β | gated | $0.0732 | $0.0732 | $0.2342 | β93% | β | β |
| Meridian AI 0x8c8cβ¦06f5 | β | 0.00 | $0.00 | 790.00 | gated | $0.0648 | $0.0648not advertised | $0.2074 | β94% | β | β |
| NovaRoute AI 0xc50dβ¦ed7b | β | 0.00 | $0.03 | 18.02M | gated | $0.0059 | $0.9286 | $0.0188 | β99% | β | β |
"Best price" and the provider table are live Antseed catalog data (advertised $/1M tokens β or $ per generated image for unit-billed image models β not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what Antseed Desktop ships with): a trust β₯ 60 gate on the effective reputation, then cheapest-first among routable providers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.