GLM-5.2 (max) logo
Z AI
•

GLM-5.2

•

Open weights model

•

Released June 2026

GLM-5.2 (max) API Provider Benchmarking & Analysis

Model Comparison

Analysis of API providers for GLM-5.2 (max) across performance metrics including latency (time to first token), output speed (output tokens per second), price and others. API providers benchmarked include Wafer, FriendliAI, Makora (NVFP4), Baseten (FAST), Novita (FP8), Baseten, Parasail (NVFP4), CoreWeave, DeepInfra (FP4), Nebius (FP4), Together AI, Scaleway, Databricks, Crusoe (NVFP4), Fireworks, GMI (FP8), and SiliconFlow (FP8).

Fastest

#1
Makora (NVFP4)Makora (NVFP4)
342.3 t/s
#2
Baseten (FAST)Baseten (FAST)
253.3 t/s
#3
Nebius (FP4)Nebius (FP4)
230.6 t/s
#4
CoreWeaveCoreWeave
205.7 t/s
#5
Together AITogether AI
205.4 t/s

Output speed

Total 17 providers

Lowest Latency

#1
Makora (NVFP4)Makora (NVFP4)
6.71 s
#2
Baseten (FAST)Baseten (FAST)
9.71 s
#3
Nebius (FP4)Nebius (FP4)
9.77 s
#4
Together AITogether AI
10.42 s
#5
CoreWeaveCoreWeave
11.11 s

Time to first answer token

Total 17 providers

Lowest Price

#1
DeepInfra (FP4)DeepInfra (FP4)
$0.49
#2
CoreWeaveCoreWeave
$0.49
#3
GMI (FP8)GMI (FP8)
$0.59
#4
Makora (NVFP4)Makora (NVFP4)
$0.66
#5
WaferWafer
$0.79

Blended price (per 1M tokens)

Total 17 providers

GLM-5.2 (max) is available through 17 API providers, each offering different performance characteristics and pricing. Below is a comparison of the key metrics across providers.

  • For output speed, the top providers are Makora (NVFP4) (342.3 t/s), Baseten (FAST) (253.3 t/s), and Nebius (FP4) (230.6 t/s). Speed varies significantly across providers, with a 461% difference between the fastest and slowest.
  • For latency, Makora (NVFP4) (6.71s), Baseten (FAST) (9.71s), and Nebius (FP4) (9.77s) offer the lowest time to first answer token.
  • For pricing, DeepInfra (FP4) (0.49), CoreWeave (0.49), and GMI (FP8) (0.59) offer the lowest blended prices per 1M tokens. Prices vary up to 5.2x across providers.
  • Makora (NVFP4) offers the best performance with both the highest speed and lowest latency. For cost optimization, DeepInfra (FP4) provides the most competitive pricing.

Highlights

Speed

Output tokens per second · Higher is better

End-to-End Response Time

Seconds · Lower is better

Price

USD per 1M tokens (blended) · Lower is better

Update: Default performance benchmarking workload has updated to 10k input tokens to better reflect production use cases. You can still select different workloads above.

Endpoint AccuracyNew

Endpoint Accuracy Index: GLM-5.2 (max)

v1.0 · Composite of BFCL v4-500, HLE-250 and AA-LCR-25 run against each provider endpoint · Percentage of the reference endpoint, with 95% confidence interval · Higher is better

Composite measure of how much of a model's accuracy a given provider endpoint preserves, from re-running BFCL v4-500, HLE-250 and AA-LCR-25 against that endpoint. Where a self-hosted reference endpoint exists, scores are expressed as a percentage of that reference (100 = matches reference); lower scores indicate accuracy lost to quantisation, sampling defaults, or other endpoint-side configuration. Scores are point-in-time snapshots. Methodology.

Pricing

Pricing: Cache Hit, Input, and Output

Price (USD per M Tokens) · Lower is better · 10,000 input tokens

Price per token for cached prompts (previously processed), typically offering a significant discount compared to regular input price, represented as USD per million tokens. The values shown here are the cache hit price; cache write and cache storage are billed separately and vary by provider — see "Cache pricing by provider" for detail.

Pricing: Blended Price

Blended at 7:2:1 (cache-input-output) · USD per 1M tokens (blended) · Lower is better

Price per token, shown in USD per million tokens. Price is a blend of cache hit, input, and output token prices using the selected ratio (default 7:2:1 cache-input-output).

Pricing: Cache Discount

1 - (cache hit price / input price) · Higher is better

Reduction in input token cost due to cache hit relative to input price. Formula: 1 - (Cache Hit Price per Token / Input Token Price), where cache hit price is the first-party cache hit price or the median provider cache hit price. Note that this discount figure does not account for all costs associated with cache hits, such as cache write and storage costs.

Output Speed vs. Price

Blended at 7:2:1 (cache-input-output) · Output speed: output tokens per second · USD per 1M tokens (blended)
Most attractive quadrant
Pareto line

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Speed

Measured by Output Speed (tokens per second)

Output Speed: GLM-5.2 (max)

Output speed: output tokens per second · 10,000 input tokens

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).

Latency vs. Output Speed

Latency: seconds to first token received · Output speed: output tokens per second · 10,000 input tokens
Most attractive quadrant
Pareto line

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Latency

Measured by Time (seconds) to First Token

Time to First Answer Token

Seconds to first token received · Lower is better · 10,000 input tokens

Time to first answer token received, in seconds, after API request sent. For reasoning models, this includes the 'thinking' time of the model before providing an answer. For models which do not support streaming, this represents time to receive the completion.

Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.

End-to-End Response Time

Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed

End-to-End Response Time

Seconds to output 500 tokens, including reasoning model 'thinking' time · Lower is better · 10,000 input tokens

Seconds to receive a 500 token response. Key components:

  • Input time: Time to receive the first response token
  • Thinking time (only for reasoning models): Time reasoning models spend outputting tokens to reason prior to providing an answer. Amount of tokens based on the average reasoning tokens across a diverse set of 60 prompts (methodology details).
  • Answer time: Time to generate 500 output tokens, based on output speed

Key Comparison Metrics & API Features

Wafer logoWafer
98%
205k
Open
$0.36
170
5.75
20.44
11.75
FriendliAI logoFriendliAI
100%
1M
Open
$0.42
196
1.29
14.04
10.20
Makora logoMakora
95%
1M
Open
$0.48
342
0.87
8.17
5.84
Baseten logoBaseten
n/a
524k
Open
$0.54
253
1.82
11.69
7.90
Novita logoNovita
98%
1.05M
Open
$0.42
79
2.13
33.78
25.32
Baseten logoBaseten
n/a
261k
Open
$0.46
110
1.95
24.70
18.20
Parasail logoParasail
98%
1M
Open
--
158
1.13
16.92
12.63
CoreWeave logoCoreWeave
97%
262k
Open
$0.38
206
1.39
13.54
9.72
DeepInfra logoDeepInfra
73%
1.05M
Open
$0.25
61
1.15
42.15
32.80
Nebius logoNebius
100%
432k
Open
$1.06
231
1.10
11.94
8.67
Together AI logoTogether AI
96%
262k
Open
$0.67
205
0.69
12.86
9.73
Scaleway logoScaleway
75%
262k
Open
$1.57
82
1.83
32.19
24.29
Databricks logoDatabricks
n/a
1M
Open
--
105
0.86
24.70
19.07
Crusoe logoCrusoe
n/a
1M
Open
$0.36
173
0.97
15.41
11.55
Fireworks logoFireworks
100%
1M
Open
$0.48
95
1.35
27.55
20.96
GMI logoGMI
n/a
1.05M
Open
--
--
--
--
--
SiliconFlow logoSiliconFlow
99%
1.05M
Open
$0.62
103
2.05
26.37
19.45

Frequently Asked Questions

Common questions about GLM-5.2 (max) providers

GLM-5.2 (max) is available through 17 API providers: Wafer, FriendliAI, Makora (NVFP4), Baseten (FAST), Novita (FP8), Baseten, Parasail (NVFP4), CoreWeave, DeepInfra (FP4), Nebius (FP4), Together AI, Scaleway, Databricks, Crusoe (NVFP4), Fireworks, GMI (FP8), and SiliconFlow (FP8). Each provider offers different performance characteristics and pricing.

GLM-5.2 (max) is currently available through 17 API providers that we benchmark and track.

The fastest providers for GLM-5.2 (max) by output speed are Makora (NVFP4) (342.3 t/s), Baseten (FAST) (253.3 t/s), and Nebius (FP4) (230.6 t/s). Output speed measures how quickly tokens are generated after the model starts responding.

The providers with the lowest time to first answer token for GLM-5.2 (max) are Makora (NVFP4) (6.71s), Baseten (FAST) (9.71s), and Nebius (FP4) (9.77s). Lower latency means faster initial response time.

The most affordable providers for GLM-5.2 (max) by blended price are DeepInfra (FP4) ($0.49 per 1M tokens), CoreWeave ($0.49 per 1M tokens), and GMI (FP8) ($0.59 per 1M tokens). Blended price uses a 7:2:1 cache hit/input/output token ratio.

The providers with the lowest input token pricing for GLM-5.2 (max) are DeepInfra (FP4) ($0.75 per 1M input tokens), CoreWeave ($0.76 per 1M input tokens), and GMI (FP8) ($0.92 per 1M input tokens).

The providers with the lowest output token pricing for GLM-5.2 (max) are DeepInfra (FP4) ($2.40 per 1M output tokens), CoreWeave ($2.42 per 1M output tokens), and GMI (FP8) ($2.90 per 1M output tokens).

Prices for GLM-5.2 (max) vary up to 5.2x across providers. The most affordable is DeepInfra (FP4) at $0.49 per 1M tokens, while Scaleway charges $2.52 per 1M tokens.

Output speed for GLM-5.2 (max) varies significantly across providers. Makora (NVFP4) is the fastest at 342.3 t/s, which is 5.6x faster than DeepInfra (FP4) at 61.0 t/s.

16 of 17 providers support JSON mode for GLM-5.2 (max): Wafer, FriendliAI, Makora (NVFP4), Baseten (FAST), Novita (FP8), Baseten, Parasail (NVFP4), CoreWeave, DeepInfra (FP4), Nebius (FP4), Together AI, Scaleway, Databricks, Crusoe (NVFP4), Fireworks, and GMI (FP8).

All 17 providers of GLM-5.2 (max) support function calling (tool use).

For GLM-5.2 (max), Makora (NVFP4) offers the best performance with highest speed and lowest latency. For cost optimization, DeepInfra (FP4) provides the most competitive pricing.

When choosing a provider for GLM-5.2 (max), consider: output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and API features like JSON mode or function calling.

Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.

For information about GLM-5.2 (max)'s intelligence, capabilities, modalities, and how it compares to other models, see the model overview page. View model overview