AI Rate CardLlama-3.1-8B-Instruct vs Llama-3.3-70B-Instruct: API price and specs
Llama-3.1-8B-Instruct is 75% cheaper than Llama-3.3-70B-Instruct on input price.
Llama-3.1-8B-Instruct is 92% cheaper than Llama-3.3-70B-Instruct on output price.
Llama-3.1-8B-Instruct is 82% cheaper than Llama-3.3-70B-Instruct on a 1M-in / 200k-out workload.
Llama-3.1-8B-Instruct has a 52.6% larger context than Llama-3.3-70B-Instruct.
| Field | Llama-3.1-8B-Instruct | Llama-3.3-70B-Instruct |
|---|---|---|
| Lowest input, per 1M | $0.02 | $0.08 +300% |
| Median input, per 1M | $0.0211 | $0.50 +2269.7% |
| Lowest output, per 1M | $0.02 | $0.26 +1180% |
| Median output, per 1M | $0.05275 | $0.72 +1264.9% |
| Lowest unit price | $0.0015 / request | — |
| Cost of 1M in + 200k out | $0.024 | $0.13 +446.7% |
| Context | 200K | 131K +34.5% |
| Max output | 131K | 131K |
| Providers | 16 +38.5% | 26 |
| Regions | EU, GLOBAL, US | EU, GLOBAL, US |
| Accepts | text | text |
| Returns | text | text |
| Type | Text models | Text models |
| Free option | no | no |
| Cheapest provider | aihubmix | requesty |
Your own workload: how many tokens the model gets (input) and gives back (output) - for example, in a month. The "Cost" row of the table above is recalculated for these numbers, for each model.
Computed from current offers each time the page opens; the best value of a row is bold, the percent is how far behind the best.
Providers selling both
aihubmix, cloudflare-workers-ai, cortecs, deepinfra, edenai, fastrouter, helicone, huggingface, jiekou, kilo, novita, openrouter, poe, requesty, trustedrouter, vercel
Frequently asked questions
Which is cheaper, Llama-3.1-8B-Instruct or Llama-3.3-70B-Instruct?
Llama-3.1-8B-Instruct is 75% cheaper than Llama-3.3-70B-Instruct on input price. Llama-3.1-8B-Instruct is 92% cheaper than Llama-3.3-70B-Instruct on output price. Llama-3.1-8B-Instruct is 82% cheaper than Llama-3.3-70B-Instruct on a 1M-in / 200k-out workload.
Which has the larger context, Llama-3.1-8B-Instruct or Llama-3.3-70B-Instruct?
Llama-3.1-8B-Instruct: 200K tokens; Llama-3.3-70B-Instruct: 131K tokens.
Where can I use both Llama-3.1-8B-Instruct and Llama-3.3-70B-Instruct?
aihubmix, cloudflare-workers-ai, cortecs, deepinfra, edenai, fastrouter, helicone, huggingface, jiekou, kilo, novita, openrouter, poe, requesty, trustedrouter, vercel