Z.ai (GLM)
GLM 5.2
GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...
Cost / 1M tokens
Medium usage$0.007
input
$0.006
cached input
$6.80
output
Context Window
1M
Up to 944K max output
Modalities
text → text
Other tokenizer
Quickstart & API
z-ai/glm-5.2Best for & Strengths
Key Strengths
Recommended Workloads in ChatOSS
- Coding assistants
- Software engineering
- Self-hosted AI
Verified Evaluations
Benchmarks & Frontier Comparison
How GLM 5.2 compares against leading open-weight and proprietary frontier models across independent agentic coding and reasoning suites.
| Benchmark (Metric) | GLM 5.2 | DeepSeek V4.1 Flash | GLM 5.3 | Kimi K3.1 Pro | Claude Opus 5.5 |
|---|---|---|---|---|---|
| Agentic & Software Engineering | |||||
| ChatOSS Agentic Coding Index (%) | 36.7% | 63.2% | 58.3% | 65% | 74.1% |
| SWE-Bench Verified (%) | 82.8% | 91.5% | 95.4% | 94.8% | 97% |
| DeepSWE v1.1 (Pass@1 %) | 43.8% | 77% | 69% | 73.5% | 74.2% |
| Vibe Code Bench v1.1 (%) | 63.96% | 87.7% | 78.1% | 88.4% | 90.3% |
| Terminal-Bench 4.0 (%) | 1.01% | 42.3% | 42% | 46.2% | 65.2% |
| Code Migration (%) | 37.87% | 45.6% | 44.2% | 52% | 66.7% |
| Intelligence & Token Economics | |||||
| Artificial Analysis Intelligence Index (0–100) | — | 39.5 | 44.8 | 55.1 | 57.6 |
| Input Price (/ 1M tokens) | $0.007 | $0.15 | $1.40 | $2.00 | $4.00 |
| Output Price (/ 1M tokens) | $6.80 | $0.60 | $4.40 | $10.00 | $20.00 |
ChatOSS Agentic Coding Index
ChatOSS Benchmark Audit · 4-benchmark composite mean
SWE-Bench Verified
Vals AI · Terminus 2
DeepSWE v1.1
DeepSWE / Official Audit · Mini-SWE-Agent
Vibe Code Bench v1.1
Vals.ai · OpenHands
Terminal-Bench 4.0
Vals.ai / Artificial Analysis · Terminus 2
Terminal-Bench 2.1
Vals.ai · Terminus 2
Code Migration
Vals.ai · Vals evaluation harness
Design Arena: AGENTICGAMEDEV
Design Arena · Rank #9 (49.7% win)
Design Arena: ANDROIDNATIVE
Design Arena · Rank #23 (52.7% win)
Design Arena: BLENDER3D
Design Arena · Rank #22 (36.1% win)
Design Arena: FULLSTACK
Design Arena · Rank #14 (60.1% win)
Active Provider Endpoints (33)
Upstream inference providers serving GLM 5.2, with per-provider pricing, context limits, quantization, and 24-hour uptime.
| Provider | Input / 1M | Output / 1M | Cache Read / 1M | Context | Max Output | Quantization | Uptime (24h) |
|---|---|---|---|---|---|---|---|
Relacecache relace/fp4 | $0.0066 | $7.98 | $0.0066 | 1M | 131K | fp4 | 99.4% |
Wafercache wafer/us | $0.007 | $6.80 | $0.006 | 1M | 944K | fp4 | 100.0% |
Wafercache wafer | $0.19 | $8.00 | $0.18 | 1M | 944K | fp4 | 100.0% |
Cloudflarecache cloudflare | $0.19 | $8.00 | $0.19 | 262K | 236K | — | 100.0% |
DeepInfracache deepinfra/fp4/flex | $0.45 | $1.44 | $0.084 | 1M | 164K | fp4 | 100.0% |
DeepInfracache deepinfra/fp4 | $0.563 | $1.80 | $0.105 | 1M | 164K | fp4 | 99.0% |
Ambientcache ambient/fp8 | $0.60 | $2.00 | $0.15 | 203K | 182K | fp8 | 85.8% |
StreamLakecache streamlake/fp8 | $0.613 | $1.93 | $0.114 | 1M | 128K | fp8 | 100.0% |
CoreWeavecache coreweave/fp4 | $0.62 | $2.42 | $0.14 | 1M | 944K | fp4 | 99.0% |
Novitacache novita/fp8 | $0.65 | $2.04 | $0.121 | 1M | 131K | fp8 | 100.0% |
Morphcache morph/fp8 | $0.699 | $2.15 | $0.236 | 1M | 944K | fp8 | 99.8% |
DigitalOceancache digitalocean | $0.70 | $2.20 | $0.105 | 1M | 131K | — | 100.0% |
SiliconFlowcache siliconflow/fp8 | $0.70 | $2.20 | $0.13 | 1M | 262K | fp8 | 99.9% |
AtlasCloudcache atlas-cloud/fp8 | $0.938 | $2.95 | $0.174 | 1M | 131K | fp8 | 98.2% |
Alibabacache alibaba/fp8 | $0.966 | $3.04 | $0.193 | 1M | 131K | fp8 | 94.6% |
Alibabacache alibaba/fp8/us-east-1 | $1.12 | $3.52 | $0.224 | 1M | 131K | fp8 | 100.0% |
Nebiuscache nebius/fp4 | $1.40 | $4.40 | $0.14 | 1M | 944K | fp4 | 32.6% |
Mistralcache mistral/zdr | $1.40 | $4.40 | $0.14 | 1M | 131K | nvfp4 | 99.6% |
BaseTencache baseten/fp8/us | $1.40 | $4.40 | $0.14 | 1M | 262K | fp8 | 98.9% |
Inceptroncache inceptron/fp4 | $1.40 | $4.40 | $0.26 | 1M | 944K | fp4 | 100.0% |
Togethercache together | $1.40 | $4.40 | $0.26 | 1M | 944K | — | 99.7% |
Baiducache baidu/fp8 | $1.40 | $4.40 | $0.26 | 1M | 131K | fp8 | 100.0% |
BaseTencache baseten/fp8 | $1.40 | $4.40 | $0.14 | 1M | 262K | fp8 | 95.4% |
Venicecache venice/fp8 | $1.40 | $4.40 | $0.26 | 1M | 131K | fp8 | 99.7% |
GMICloudcache gmicloud/fp8 | $1.40 | $4.40 | $0.26 | 1M | 944K | fp8 | 98.5% |
Parasailcache parasail/fp4 | $1.40 | $4.40 | $0.26 | 262K | 236K | fp4 | 96.5% |
Friendlicache friendli | $1.40 | $4.40 | $0.26 | 1M | 944K | — | 98.5% |
Z.AIcache z-ai/fp8 | $1.40 | $4.40 | $0.26 | 1M | 131K | fp8 | 100.0% |
Mistralcache mistral/eu | $1.54 | $4.84 | $0.154 | 1M | 128K | — | 99.3% |
BaseTencache baseten/us/fast | $2.10 | $6.60 | $0.21 | 1M | 262K | fp8 | 97.8% |
BaseTencache baseten/fast | $2.10 | $6.60 | $0.21 | 1M | 262K | fp8 | 97.8% |
Baiducache baidu/fast | $2.25 | $7.88 | $0.56 | 1M | 131K | fp4 | 100.0% |
Alibabacache alibaba/fast | $2.31 | $7.26 | $0.462 | 1M | 131K | fp8 | 92.7% |
Supported API Parameters
Parameters accepted by this model when called through ChatOSS or OpenRouter.
Deep-Dive Benchmarks & Teardowns
Independent benchmark analysis and real-world agent billing teardowns featuring GLM 5.2.
Compare Similar Models
Browse all models →Z.ai (GLM)
GLM 5.3 FlashX Legal
GLM-5.3-FlashX-Legal (GLM-Legal) is Z.ai's legal-domain variant of GLM-5.3-FlashX, a native multimodal model that accepts text, image, and video input and produces text. It adds legal-domain mid-training and post-training with reinforcement...
Z.ai (GLM)
GLM 5.3 Prime
GLM-5.3-Prime is the high-speed variant of Z.ai's GLM-5.3, inheriting its full capabilities while delivering 1.5–2× the output throughput through inference acceleration. It supports text input and output with a 1M-token...
Z.ai (GLM)
GLM 5.3 FlashX
GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...
Z.ai (GLM)
GLM 5.3 Flash
GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...
Anthropic
Claude Opus 5.5
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...
Moonshot AI
Kimi K3.1 Pro
Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, with a 1M-token context window. It is built for long-horizon coding, 3D...