Granite 4.0 H Micro vs Llama 3 8B Lunaris
Prices checked
Granite 4.0 H Micro takes 3 of 5 rounds against Llama 3 8B Lunaris. It is also the cheaper of the two to run, at $0.04 per million tokens against $0.04.
Granite 4.0 H Micro
131k context · tool calling · open weights
$0.04/M tokensCloudflare Workers AI's own price · 1.7¢ per 1,000 pages
Llama 3 8B Lunaris
8k context · open weights
$0.04/M tokensOpenRouter's own price · 2.7¢ per 1,000 pages
Maker's price in (lower wins)
$0.02/M
$0.04/M
Maker's price out (lower wins)
$0.11/M
$0.05/M
Cheapest price (lower wins)
none below the maker
none below the maker
Maker's price, blended (lower wins)
$0.04/M
$0.04/M
Cost per 1,000 pages (lower wins)
1.7¢
2.7¢
Context
131k tokens
8k tokens
Open weights
yes
yes
Tool calling
yes
no
Structured output
yes
yes
Reads images
no
no
Reasoning
no
no
Released
2 Oct 2025
13 Aug 2024
Providers selling it
4
5
Price we measure against
Cloudflare Workers AI
OpenRouter
Trains on your prompts (no is better)
no provider to check
no provider to check
Providers selling both · 1
- AI/ML API$0.04 / $0.05
Granite 4.0 H Micro, then Llama 3 8B Lunaris, per million tokens.
Go further
Add a third or fourth model, or change the weighting, in the tool.
Frequently asked
- Which is cheaper, Granite 4.0 H Micro or Llama 3 8B Lunaris?
- Granite 4.0 H Micro, at $0.04 per million tokens against $0.04 for Llama 3 8B Lunaris, blended 3 in : 1 out, read on 31 Oct 2026.
- Where is each one cheapest?
- Granite 4.0 H Micro: nobody sells it below Cloudflare Workers AI's own price with a stated reason, so $0.04 is the number to beat. Llama 3 8B Lunaris: nobody sells it below OpenRouter's own price with a stated reason, so $0.04 is the number to beat.
- Can I buy both from one provider?
- Yes. One provider sells both: AI/ML API. The cheapest of them for the two together is AI/ML API.
- Which holds more context?
- Granite 4.0 H Micro, at 131k tokens against 8k.
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