GLM-5.1 vs GPT-5.6 Terra Pro
Prices checked
GLM-5.1 takes 3 of 5 rounds against GPT-5.6 Terra Pro. It is also the cheaper of the two to run, at $1.08 per million tokens against $2.25.
GLM-5.1
200k context · tool calling · open weights
$1.08/M tokensCheapest price, at HPC-AI · $0.52 per 1,000 pages
GPT-5.6 Terra Pro
1.1M context · tool calling · reads images
$2.25/M tokensOpenAI's own price · $0.96 per 1,000 pages
Maker's price in (lower wins)
$1.40/M
$1/M
Maker's price out (lower wins)
$4.40/M
$6/M
Cheapest price (lower wins)
$1.08/M at HPC-AI
none below the maker
Maker's price, blended (lower wins)
$2.15/M
$2.25/M
Cost per 1,000 pages (lower wins)
$0.52
$0.96
Context
200k tokens
1.1M tokens
Open weights
yes
not published
Tool calling
yes
yes
Structured output
yes
yes
Reads images
no
yes
Reasoning
yes
yes
Released
7 Apr 2026
9 Jul 2026
Providers selling it
58
10
Price we measure against
the maker's own
the maker's own
Trains on your prompts (no is better)
not published
no provider to check
Providers selling both · 6
NanoGPT$1.15 / $4.50
Nous Portal$1.40 / $4.50
LLMTR$1.94 / $4.50
302.AI$2.15 / $4.50
Kilo Gateway$2.15 / $4.50
OpenRouter$2.15 / $4.50
GLM-5.1, then GPT-5.6 Terra Pro, per million tokens.
Go further
Add a third or fourth model, or change the weighting, in the tool.
What moved this week
Frequently asked
- Which is cheaper, GLM-5.1 or GPT-5.6 Terra Pro?
- GLM-5.1, at $1.08 per million tokens against $2.25 for GPT-5.6 Terra Pro, blended 3 in : 1 out, read on 31 Oct 2026.
- Where is each one cheapest?
- GLM-5.1: $1.08 at HPC-AI. GPT-5.6 Terra Pro: nobody sells it below the maker's own price with a stated reason, so $2.25 is the number to beat.
- Can I buy both from one provider?
- Yes. 6 providers sell both: NanoGPT, Nous Portal, LLMTR, 302.AI, Kilo Gateway, OpenRouter. The cheapest of them for the two together is NanoGPT.
- Which holds more context?
- GPT-5.6 Terra Pro, at 1.1M tokens against 200k.
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