GLM 5.3 vs GLM 5.2 API pricing
Updated
GLM 5.3: USD 0.5625 input, USD 2.50 output per 1M tokens (DeepInfra, promotional). GLM 5.2: USD 0.5625 input, USD 1.80 output per 1M tokens (DeepInfra, promotional). At 3 input tokens per output token, GLM 5.3 costs 20% more than GLM 5.2 (USD 1.047 vs USD 0.8719 per 1M).
How do GLM 5.3 and GLM 5.2 differ on price?
- Cheapest input price: GLM 5.3 USD 0.19 against USD 0.5625 per 1M tokens, so GLM 5.2 costs 3× as much.
- Cheapest output price: GLM 5.2 USD 1.80 against USD 2.20 per 1M tokens, so GLM 5.3 costs 1.22× as much (22% more).
- Among the 201 indexed models, GLM 5.3 ranks 133rd cheapest and GLM 5.2 123rd on the 3:1 blend.
- The cheaper model changes at 7 input tokens per output token.
- Both list cached-input discounts: GLM 5.3 at 29 standard prices of 31, from USD 0.075 (Relace, 90% below that provider's input price); GLM 5.2 at 21 standard prices of 21, from USD 0.075 (Relace, 90% below that provider's input price).
- Both state a context window of up to 1,048,576 tokens.
- GLM 5.3 has 31 providers with a standard price, 10 more than GLM 5.2; 17 providers sell both (DeepInfra, Relace, GMI Cloud, Doubleword and 13 more).
- On one account, DeepInfra charges USD 1.047 for GLM 5.3 (its cheapest price) and USD 0.8719 for GLM 5.2 (its cheapest price) at 3:1, the lowest combined price among the providers that sell both.
- Shopping around matters more for GLM 5.3: its dearest standard price is 4.1× its cheapest, against 2.7× for GLM 5.2.
- Z.ai's own API charges USD 2.15 for GLM 5.3 at 3:1, 105% above the cheapest provider (DeepInfra).
- Z.ai's own API charges USD 2.15 for GLM 5.2 at 3:1, 147% above the cheapest provider (DeepInfra).
- GLM 5.3 has a batch price at 2 providers, from USD 1.075 at 3:1 (Doubleword); GLM 5.2 has a batch price at 1 provider, from USD 0.7275 at 3:1 (Doubleword).
- For the one-developer coding agent preset, the gap is USD 704.70 a year in GLM 5.2's favour (GLM 5.3 USD 533.25 a month, GLM 5.2 USD 474.53).
- Per character of input, GLM 5.3 costs 1.21× as much as GLM 5.2 in every measured language, because both use the GLM 5.3 tokenizer.
GLM 5.3 and GLM 5.2 prices side by side
USD per 1M tokens, standard tier. "Cheapest" is the lowest price at 3 input tokens to 1 output token; each line can come from a different provider.
| Price | GLM 5.3 | GLM 5.2 |
|---|---|---|
| Model page | GLM 5.3 Z.ai · 31 providers | GLM 5.2 Z.ai · 21 providers |
| Cheapest standard price, blended 3:1 | USD 1.047 promo USD 0.5625 in / USD 2.50 out at DeepInfra DeepInfra page checked 2026-11-14 | USD 0.8719 lower promo USD 0.5625 in / USD 1.80 out at DeepInfra DeepInfra page checked 2026-11-14 |
| Cheapest input | USD 0.19 lower promo AkashML source | USD 0.5625 promo DeepInfra source |
| Cheapest cached input | USD 0.075 tie Relace source | USD 0.075 tie Relace source |
| Cheapest output | USD 2.20 promo GMI Cloud source | USD 1.80 lower promo DeepInfra source |
| Cheapest non-promotional price | USD 0.75 / USD 2.40 Relace, USD 1.163 blended source | USD 0.62 / USD 2.42 CoreWeave Forge, USD 1.07 blended source |
| Maker's own API | USD 1.40 / USD 4.40 Z.ai source | USD 1.40 / USD 4.40 Z.ai source |
| Median standard price | USD 1.40 / USD 4.40 input / output, 31 prices | USD 1.40 / USD 4.40 input / output, 21 prices |
| Context window | 1,048,576 tokens, largest stated; from 262,144 at some providers | 1,048,576 tokens, largest stated; from 262,144 at some providers |
| Providers with a standard price | 31 | 21 |
- Lower marks the lower price on a line; tie, equal prices. Other tiers and free rows are on the model pages.
- Promo: the provider calls the price promotional or discounted, so it is current now and can end without notice.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
What do GLM 5.3 and GLM 5.2 cost for a real workload?
30 days at each model's cheapest standard price for the workload (presets of the cost calculator; cached input at the cached price where listed).
| Workload | GLM 5.3 | GLM 5.2 | Difference |
|---|---|---|---|
| Support chatbot 2,000 requests a day, 1,500 input tokens (60% cached), 300 output | USD 71.82 GMI Cloud | USD 58.32 DeepInfra | USD 13.50 GLM 5.2 lower |
| RAG over documents 1,000 requests a day, 8,000 input tokens (30% cached), 500 output | USD 111.60 AkashML | USD 129.06 DeepInfra | USD 17.46 GLM 5.3 lower |
| Coding agent (one developer) 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning | USD 533.25 Relace | USD 474.53 DeepInfra | USD 58.72 GLM 5.2 lower |
The cheaper model depends on the workload: GLM 5.2 for Support chatbot, GLM 5.3 for RAG over documents, GLM 5.2 for Coding agent (one developer). Cache-heavy work rewards a low cached-input price. The widest gap is on Support chatbot: the dearer model costs 23% more than the cheaper one.
Does the cheaper model change with your input:output ratio?
The cheaper model changes with the mix: below 7 input tokens per output token, GLM 5.2 is cheaper; above 7, GLM 5.3 is cheaper. At the site's 3:1 blend, GLM 5.2 is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.
| Input : output tokens | GLM 5.3 | GLM 5.2 |
|---|---|---|
| 1 : 3 (long answers) | USD 1.825 GMI Cloud | USD 1.491 lower DeepInfra |
| 1 : 1 | USD 1.45 GMI Cloud | USD 1.181 lower DeepInfra |
| 3 : 1 (the site’s blend) | USD 1.047 DeepInfra | USD 0.8719 lower DeepInfra |
| 10 : 1 (retrieval, long prompts) | USD 0.5727 lower AkashML | USD 0.675 DeepInfra |
| 30 : 1 (agents, big contexts) | USD 0.3258 lower AkashML | USD 0.6024 DeepInfra |
USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.
What do GLM 5.3 and GLM 5.2 cost per character in other languages?
Cost to send 1M characters at each model's cheapest non-promotional standard input price, with the GLM 5.3 tokenizer for GLM 5.3 (the same one for both, so the gap equals the price gap):
| Content | GLM 5.3 | GLM 5.2 |
|---|---|---|
| English | USD 0.163 Relace | USD 0.135 lower CoreWeave Forge |
| Arabic MSA | USD 0.310 Relace | USD 0.256 lower CoreWeave Forge |
| Egyptian Arabic | USD 0.314 Relace | USD 0.260 lower CoreWeave Forge |
| Hindi | USD 0.671 Relace | USD 0.555 lower CoreWeave Forge |
| Chinese | USD 0.445 Relace | USD 0.368 lower CoreWeave Forge |
| Spanish | USD 0.187 Relace | USD 0.155 lower CoreWeave Forge |
| Code | USD 0.191 Relace | USD 0.158 lower CoreWeave Forge |
From the real cost by language dataset (measured tokens per character on FLORES-200 and other texts). Your own text will differ.
Published results on the same benchmark
BroadNet did not run these benchmarks. BroadRouter is BroadNet's site. Each row pairs results from one source with the same setting for both models, as published; results from different sources or settings are never put side by side. One benchmark says little about your own workload.
| Benchmark | GLM 5.3 | GLM 5.2 | Source and setting |
|---|---|---|---|
| GPQA Diamond | 0.909 | 0.919 | Epoch AI, checked 2026-11-16 Max reasoning setting, mean score on a 0 to 1 scale |
Epoch AI rows: Epoch AI, Capabilities & benchmarking, licensed CC BY 4.0, shown as the mean score on a 0 to 1 scale, rounded to three decimals (hover a value for the exact figure).
All published results for each model, with every setting: GLM 5.3 · GLM 5.2.
GLM 5.3 on BroadNet Inference
BroadRouter is BroadNet's own site, so BroadNet's price is shown here for reference and is not part of the comparison above. BroadNet Inference serves GLM 5.3 as broadnet/glm-5.3 at USD 0.78 input, USD 0.173 cached and USD 3.46 output per 1M tokens (USD 1.45 blended 3:1; the cheapest third-party standard price above is USD 1.047). On BroadNet it reads up to 1,048,576 tokens per request and accepts text input (BroadNet rate card).
Check it on your own traffic
- GLM 5.3 API pricing: all 31 providers and other tiers.
- GLM 5.2 API pricing: all 21 providers and other tiers.
- Cost calculator with GLM 5.3 and GLM 5.2 preloaded
- GLM token counter
- Prompt caching calculator
More comparisons: DeepSeek V4 Pro vs GLM 5.3 · GLM 5.3 vs MiniMax M3 · GLM 5.3 vs Qwen3.8 27B · GLM 5.3 vs Kimi K3 · Claude Sonnet 5.5 vs GLM 5.3 · DeepSeek V4.1 Flash vs GLM 5.3 · GLM 5.3 vs Nemotron 3 Ultra · GLM 5.3 vs Kimi K2.5 · GLM 5.3 vs DeepSeek V4 Pro 0813 · GLM 5.3 vs Kimi K2.6 · GLM 5.2 vs GLM 5.1 · every GLM 5.3 comparison · every GLM 5.2 comparison · all 662 price comparisons
Questions
Is GLM 5.3 cheaper than GLM 5.2?
No. On a blend of 3 input tokens to 1 output token, GLM 5.2 costs USD 0.8719 per 1M tokens and GLM 5.3 USD 1.047, at each model's cheapest standard price. For input alone, GLM 5.3 is cheaper (USD 0.19 against USD 0.5625 per 1M); for output alone, GLM 5.2 is (USD 1.80 against USD 2.20). This compares price only, not quality.
What do GLM 5.3 and GLM 5.2 cost per 1M tokens?
GLM 5.3: from USD 0.5625 input and USD 2.50 output at DeepInfra (promotional); Z.ai's own API charges USD 1.40 / USD 4.40; median of 31 standard prices USD 1.40 / USD 4.40. GLM 5.2: from USD 0.5625 input and USD 1.80 output at DeepInfra (promotional); Z.ai's own API charges USD 1.40 / USD 4.40; median of 21 standard prices USD 1.40 / USD 4.40. Each price is read from the provider's own pricing page; the source is linked in the table.
Which providers sell both GLM 5.3 and GLM 5.2?
17 providers have a standard price for both: DeepInfra (USD 1.047 and USD 0.8719 at 3:1), Relace (USD 1.163 and USD 1.163 at 3:1), GMI Cloud (USD 1.075 and USD 2.15 at 3:1), Doubleword (USD 2.15 and USD 1.448 at 3:1), Chutes (USD 1.881 and USD 1.925 at 3:1), FriendliAI (USD 1.935 and USD 2.15 at 3:1), SiliconFlow (USD 2.15 and USD 1.999 at 3:1), Alibaba Cloud (USD 2.15 and USD 2.15 at 3:1), Baseten (USD 2.15 and USD 2.15 at 3:1), Cloudflare Workers AI (USD 2.15 and USD 2.15 at 3:1), DigitalOcean (USD 2.15 and USD 2.15 at 3:1), Featherless (USD 2.15 and USD 2.15 at 3:1), Novita AI (USD 2.15 and USD 2.15 at 3:1), Parasail (USD 2.15 and USD 2.15 at 3:1), Together AI (USD 2.15 and USD 2.15 at 3:1), Z.ai (USD 2.15 and USD 2.15 at 3:1), Venice (USD 2.688 and USD 2.15 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 1.047 and USD 0.8719.
Which is cheaper for a coding agent or a chatbot, GLM 5.3 or GLM 5.2?
For a coding agent (1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning): USD 533.25 a month on GLM 5.3 (Relace) against USD 474.53 on GLM 5.2 (DeepInfra), a difference of USD 58.72. For a support chatbot (2,000 requests a day, 1,500 input tokens (60% cached), 300 output): USD 71.82 a month on GLM 5.3 (GMI Cloud) against USD 58.32 on GLM 5.2 (DeepInfra), a difference of USD 13.50. Each uses the cheapest standard price for that workload, cached input included. Try your own numbers.
At what input:output ratio does the cheaper of GLM 5.3 and GLM 5.2 change?
The cheaper model changes with the mix: below 7 input tokens per output token, GLM 5.2 is cheaper; above 7, GLM 5.3 is cheaper. At the site's 3:1 blend, GLM 5.2 is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.
Do GLM 5.3 and GLM 5.2 have cached input pricing?
GLM 5.3 from USD 0.075 per 1M cached input tokens (Relace; 29 of 31 standard prices list a discount). GLM 5.2 from USD 0.075 per 1M cached input tokens (Relace; 21 of 21 standard prices list a discount). Cached tokens are billed at the input price where no cache price is published.
Which has the larger context window, GLM 5.3 or GLM 5.2?
Both go up to 1,048,576 tokens on the providers that state it.
Can I use GLM 5.3 on BroadNet Inference?
BroadNet Inference serves GLM 5.3 as broadnet/glm-5.3 at USD 0.78 input and USD 3.46 output per 1M tokens (rate card synced 2026-11-14). BroadRouter is BroadNet's own site, so BroadNet's price is shown for reference and not ranked against the providers above.
Which scores higher on GPQA Diamond, GLM 5.3 or GLM 5.2?
Per Epoch AI (checked 2026-11-16), GLM 5.3 scores 0.909 and GLM 5.2 scores 0.919 on GPQA Diamond (Max reasoning setting, mean score on a 0 to 1 scale). BroadNet did not run these benchmarks, and one benchmark says little about your own workload. Every result both models share, with sources.
Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity
Sources
- DeepInfra: pricing (checked 2026-11-14)
- AkashML: pricing (checked 2026-11-15)
- GMI Cloud: pricing (checked 2026-11-15)
- Relace: pricing (checked 2026-11-16)
- Z.ai: own API pricing (checked 2026-11-14)
- DeepInfra: pricing (checked 2026-11-14)
- CoreWeave Forge: pricing (checked 2026-11-15)
- GMI Cloud: pricing (checked 2026-11-15)
- Doubleword: pricing (checked 2026-11-16)
- Chutes: pricing (checked 2026-11-15)
- FriendliAI: pricing (checked 2026-11-15)
- FriendliAI: pricing (checked 2026-11-15)
- SiliconFlow: pricing (checked 2026-11-14)
- Alibaba Cloud: pricing (checked 2026-11-14)
- Baseten: pricing (checked 2026-11-14)
- Cloudflare Workers AI: pricing (checked 2026-11-14)
- DigitalOcean: pricing (checked 2026-11-14)
- Featherless: pricing (checked 2026-11-15)
- Featherless: pricing (checked 2026-11-15)
- Novita AI: pricing (checked 2026-11-14)
- Parasail: pricing (checked 2026-11-14)
- Together AI: pricing (checked 2026-11-14)
- Venice: pricing (checked 2026-11-15)
- Epoch AI benchmark data, CC BY 4.0 (checked 2026-11-16)
- BroadNet Inference rate card, pricing.json (synced 2026-11-14)