Grok 4.20 vs Grok 4.3
Prices checked
Grok 4.20 takes 1 of 5 rounds against Grok 4.3. Both cost $1.56 per million tokens to run.
xAIWins 1 roundGrok 4.20
2M context · tool calling · reads images
$1.56/M tokensxAI's own price · $0.90 per 1,000 pages
xAIGrok 4.3
1M context · tool calling · reads images
$1.56/M tokensxAI's own price · $0.90 per 1,000 pages
Maker's price in (lower wins)
$1.25/M
$1.25/M
Maker's price out (lower wins)
$2.50/M
$2.50/M
Cheapest price (lower wins)
none below the maker
none below the maker
Maker's price, blended (lower wins)
$1.56/M
$1.56/M
Cost per 1,000 pages (lower wins)
$0.90
$0.90
Context
2M tokens
1M tokens
Open weights
not published
no
Tool calling
yes
yes
Structured output
yes
yes
Reads images
yes
yes
Reasoning
yes
yes
Released
31 Mar 2026
17 Apr 2026
Providers selling it
9
34
Price we measure against
the maker's own
the maker's own
Trains on your prompts (no is better)
no provider to check
no provider to check
Providers selling both · 9
Eden AI$1.56 / $1.56
Kilo Gateway$1.56 / $1.56
NanoGPT$1.56 / $1.56
Nous Portal$1.56 / $1.56
OpenRouter$1.56 / $1.56- xAI$1.56 / $1.56
TrustedRouter$1.65 / $1.65
Venice$1.77 / $1.77
Grok 4.20, then Grok 4.3, per million tokens; the 8 cheapest of 9.
Go further
Add a third or fourth model, or change the weighting, in the tool.
Frequently asked
- Which is cheaper, Grok 4.20 or Grok 4.3?
- Neither: both run at $1.56 per million tokens, blended 3 in : 1 out, read on 31 Oct 2026.
- Where is each one cheapest?
- Grok 4.20: nobody sells it below the maker's own price with a stated reason, so $1.56 is the number to beat. Grok 4.3: nobody sells it below the maker's own price with a stated reason, so $1.56 is the number to beat.
- Can I buy both from one provider?
- Yes. 9 providers sell both: Eden AI, Kilo Gateway, NanoGPT, Nous Portal, OpenRouter, xAI, and more. The cheapest of them for the two together is Eden AI.
- Which holds more context?
- Grok 4.20, at 2M tokens against 1M.
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