DeepSeek V3.1 vs GLM-4.5-Air
Prices checked
DeepSeek V3.1 takes 3 of 5 rounds against GLM-4.5-Air. GLM-4.5-Air counters on price: $0.16 per million tokens against $0.32, for 131k context instead of 164k.
DeepSeekWins 3 roundsDeepSeek V3.1
164k context · tool calling · open weights
$0.32/M tokensCheapest price, at TokenGo · $0.16 per 1,000 pages
Z.aiGLM-4.5-Air
131k context · tool calling · open weights
$0.16/M tokensCheapest price, at ZenMux · 8.7¢ per 1,000 pages
Maker's price in (lower wins)
$0.25/M
$0.20/M
Maker's price out (lower wins)
$0.95/M
$1.10/M
Cheapest price (lower wins)
$0.32/M at TokenGo
$0.16/M at ZenMux
Maker's price, blended (lower wins)
$0.42/M
$0.43/M
Cost per 1,000 pages (lower wins)
$0.16
8.7¢
Context
164k tokens
131k tokens
Open weights
yes
yes
Tool calling
yes
yes
Structured output
yes
not published
Reads images
no
no
Reasoning
yes
yes
Released
21 Aug 2025
28 Jul 2025
Providers selling it
30
24
Price we measure against
OpenRouter
the maker's own
Trains on your prompts (no is better)
not published
not published
Providers selling both · 14
- Submodel$0.35 / $0.20
NanoGPT$0.33 / $0.31
Kilo Gateway$0.42 / $0.31
Nous Portal$0.42 / $0.31
OpenRouter$0.42 / $0.31
Eden AI$0.42 / $0.32
Hugging Face Inference Providers$0.45 / $0.31
Novita$0.45 / $0.31
DeepSeek V3.1, then GLM-4.5-Air, per million tokens; the 8 cheapest of 14.
Go further
Add a third or fourth model, or change the weighting, in the tool.
What moved this week
- NanoGPT · 7% dearer · $0.13 / $0.85 $0.12 / $0.80
- NanoGPT · 7% dearer · $0.13 / $0.85 $0.12 / $0.80
Frequently asked
- Which is cheaper, DeepSeek V3.1 or GLM-4.5-Air?
- GLM-4.5-Air, at $0.16 per million tokens against $0.32 for DeepSeek V3.1, blended 3 in : 1 out, read on 31 Oct 2026.
- Where is each one cheapest?
- DeepSeek V3.1: $0.32 at TokenGo. GLM-4.5-Air: $0.16 at ZenMux.
- Can I buy both from one provider?
- Yes. 14 providers sell both: Submodel, NanoGPT, Kilo Gateway, Nous Portal, OpenRouter, Eden AI, and more. The cheapest of them for the two together is Submodel.
- Which holds more context?
- DeepSeek V3.1, at 164k tokens against 131k.
Prices exclude tax, read 31 Oct 2026 from each provider's own catalog. They are per M tokens, blended 3 in : 1 out, and a page is 600 tokens in and 60 out. Bars are scaled to the larger value in each row.
Related comparisons