Z.ai: GLM 5.3 (batch)
z-ai/glm-5-3-batchGLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.
Modalities
In / out per 1M$0.45 / $2.00
Context1M
ReleasedAug 18, 2026
Providers
46 providers serve this model upstream.
| Provider | Input /M | Output /M | Cache read /M | Latency P50 | Throughput P50 | Uptime |
|---|---|---|---|---|---|---|
| $0.08 | $1.40 | $0.08 | 0.49s | 119 tps | 100% | |
| $0.18 | $1.40 | $0.18 | 0.44s | 109 tps | 100% | |
| $0.50 | $1.50 | $0.12 | 1.71s | 98 tps | 100% | |
| $0.50 | $1.50 | $0.12 | 1.41s | 84 tps | 100% | |
| $0.75 | $2.40 | $0.07 | 1.91s | 83 tps | 100% | |
| $0.84 | $2.68 | $0.19 | 0.85s | 95 tps | 100% | |
| $1.40 | $4.40 | $0.26 | 0.38s | 139 tps | 100% | |
| $0.19 | $4.40 | $0.19 | 1.59s | 71 tps | 100% | |
| $1.40 | $4.40 | $0.26 | 1.05s | 122 tps | 100% | |
| $0.84 | $2.64 | $0.16 | 2.65s | 43 tps | 100% | |
| $1.40 | $4.40 | $0.26 | 3.19s | 48 tps | 100% | |
| $0.70 | $2.20 | $0.13 | 2.43s | 39 tps | 100% | |
| $1.19 | $3.74 | $0.24 | 1.79s | 55 tps | 100% | |
| $2.80 | $8.80 | $0.56 | 1.32s | 59 tps | 100% | |
| $1.40 | $4.40 | $0.26 | 6.37s | 30 tps | 100% | |
 Mistraleu · nvfp4 | $1.54 | $4.84 | $0.15 | 3.09s | 106 tps | 100% |
 Baseten (fp8)us · fp8 | $2.10 | $6.60 | $0.21 | 0.75s | 149 tps | 100% |
| Baseten (fp8)fp8 | $2.10 | $6.60 | $0.21 | 0.74s | 140 tps | 100% |
| Baseten (fp4)fp4 | $1.40 | $4.40 | $0.14 | 5.79s | 41 tps | 100% |
| $0.08 | $4.38 | $0.08 | 0.85s | 110 tps | 99.97% | |
| $0.55 | $1.40 | $0.18 | 0.81s | 87 tps | 99.97% | |
| $1.40 | $4.40 | $0.26 | 1.49s | 50 tps | 99.96% | |
| $0.49 | $2.60 | $0.19 | 42.12s | 3 tps | 99.91% | |
| $0.04 | $6.00 | $0.03 | 1.25s | 37 tps | 99.89% | |
 SiliconFlowfp8 | $0.70 | $2.20 | $0.13 | 2.86s | 58 tps | 99.88% |
| Baseten (fp4)us · fp4 | $1.40 | $4.40 | $0.14 | 3.65s | 42 tps | 99.87% |
 Fireworks | $1.40 | $4.40 | $0.26 | 2.41s | 53 tps | 99.86% |
 DigitalOcean | $0.91 | $2.86 | $0.17 | 0.96s | 89 tps | 99.83% |
| $1.40 | $4.40 | $0.26 | 1.64s | 122 tps | 99.82% | |
| $1.40 | $4.40 | $0.26 | 2.67s | 40 tps | 99.76% | |
| Fireworks | $2.10 | $6.60 | $0.39 | 0.96s | 144 tps | 99.74% |
| Mistralnvfp4 | $1.40 | $4.40 | $0.14 | 4.20s | 34 tps | 99.68% |
| $1.40 | $4.40 | $0.26 | 0.74s | 109 tps | 99.57% | |
| $0.14 | $4.20 | $0.10 | 3.17s | 13 tps | 99.51% | |
| $1.26 | $3.96 | $0.23 | 0.78s | 93 tps | 99.47% | |
 DeepInfraus · fp4 | $0.56 | $2.50 | $0.13 | 2.21s | 40 tps | 99.26% |
| $1.40 | $4.40 | $0.26 | 0.89s | 79 tps | 99.19% | |
| Mistralnvfp4 | $1.40 | $4.40 | $0.14 | 13.50s | 96 tps | 99.07% |
| $0.18 | $2.95 | $0.15 | 1.34s | 63 tps | 98.68% | |
 ModelRun [by Modular]fp4 | $2.10 | $6.60 | $0.40 | 0.32s | 153 tps | 98.25% |
| $1.40 | $4.40 | $0.26 | 1.55s | 60 tps | 98.22% | |
| $1.40 | $4.40 | $0.26 | 2.29s | 24 tps | 98.05% | |
| $0.14 | $4.40 | $0.10 | 1.13s | 75 tps | 97.75% | |
| $1.40 | $4.40 | Variable | — | — | 94.93% | |
 CoreWeavefp8 | $1.20 | $4.00 | $0.19 | 0.84s | 99 tps | 92.54% |
| $0.60 | $2.50 | $0.10 | 2.01s | 72 tps | 87.41% |
Market-wide measurements, refreshed every 15 minutes. They describe who serves GLM 5.3 (batch) in general, not how ProjectCOZY routes your requests.
Pricing
List rates per 1M tokens — what GLM 5.3 (batch) costs pay-as-you-go elsewhere, not what you pay here.
Input
$0.45
per 1M tokens
Output
$2.00
per 1M tokens
Cache read
$0.10
per 1M cached tokens
Included in every pass
GLM 5.3 (batch) is included in every ProjectCOZY pass at no extra cost — one flat rate, no per-token metering. The rates above are there for comparison.
Performance
Time to first token and output speed across every provider, over the last 3 days.
0.0s100.0s200.0s300.0s400.0s
Oct 16, 12 AMOct 16, 8 AMOct 16, 4 PMOct 17, 12 AMOct 17, 8 AMOct 17, 4 PMOct 18, 12 AMOct 18, 8 AMOct 18, 4 PMOct 19, 12 AMOct 19, 8 AM
Prime Intellect Together GMICloud Alibaba Cloud Int. Cloudflare Baseten (fp8) (us) SiliconFlow Baidu Qianfan Makora DeepInfra (us) Sail Research Phala Fireworks AkashML ModelRun [by Modular] Wafer Alibaba Cloud Int. inference.net Mistral NovitaAI Mistral (eu) Tenstorrent AtlasCloud Sail Research (us) Mistral Reka AI Z.ai Baseten (fp8) Relace io.net Friendli Inceptron (eu) Fireworks Wafer (us) Modal Baseten (fp4) (us) Venice Decart Crusoe Parasail Morph DigitalOcean Baseten (fp4) DekaLLM Nebius Token Factory CoreWeave
0 tok/s2000 tok/s4000 tok/s6000 tok/s8000 tok/s
Oct 16, 12 AMOct 16, 8 AMOct 16, 4 PMOct 17, 12 AMOct 17, 8 AMOct 17, 4 PMOct 18, 12 AMOct 18, 8 AMOct 18, 4 PMOct 19, 12 AMOct 19, 8 AM
Prime Intellect Together GMICloud Alibaba Cloud Int. Cloudflare Baseten (fp8) (us) SiliconFlow Baidu Qianfan Makora DeepInfra (us) Sail Research Phala Fireworks AkashML ModelRun [by Modular] Wafer Alibaba Cloud Int. inference.net Mistral NovitaAI Mistral (eu) Tenstorrent AtlasCloud Sail Research (us) Mistral Reka AI Z.ai Baseten (fp8) Relace io.net Friendli Inceptron (eu) Fireworks Wafer (us) Modal Baseten (fp4) (us) Venice Decart Crusoe Parasail Morph DigitalOcean Baseten (fp4) DekaLLM Nebius Token Factory CoreWeave
Uptime
100% of requests were served successfully over the last 3 days, across all providers.
Last 3 daysAvailability 99.9%
Oct 16, 12:00 PMNow
Availability over the last 24 hoursAvailability 99.93%
100% 98% 97% 95%
Oct 18, 12:10 PMNow
Apps
Tools that most often run GLM 5.3 (batch), by tokens this week.
Activity
15682.6B tokens processed market-wide in the last month.
0250.0B500.0B750.0B1000.0B
Sep 19Sep 24Sep 29Oct 4Oct 9Oct 14Oct 19
Tokens
FAQ
Which model id do I use for GLM 5.3 (batch)?
z-ai/glm-5.3:batch. Send it as the model field from any OpenAI-compatible client, against http://192.168.1.55/v1.
How much context does GLM 5.3 (batch) hold?
1,048,576 tokens per request, of which up to 131,072 can be the response.
What can it accept and return?
Input: text. Output: text.
What does GLM 5.3 (batch) cost on a pass?
Nothing beyond the pass itself. Pay-as-you-go elsewhere it lists at $0.45 per 1M input tokens and $2.00 per 1M output tokens.
Who built GLM 5.3 (batch)?
Z.ai, in the GLM family.
Explore
Other models your pass can call.
