InfraRabbit

Llama 3.3 70B Instruct API price

The cheapest host lists Llama 3.3 70B Instruct at USD 0.10 per 1M input tokens and USD 0.32 output (DeepInfra, fp8 quantization, inference/openrouter/meta-llama/llama-3.3-70b-instruct/deepinfra/fp8); the median across hosts is USD 0.4415. 10 hosts, 2026-10-24 (UTC). Hosts run different quantizations; each row states its own.

List prices, not transacted prices. What that means.

Cheapest Llama 3.3 70B Instruct input price

1 line, 2026-10-19 to 2026-10-24; values between 0.1 and 0.1 USD; latest 0.1 USD.0.090.0950.10.1050.11Oct 19Oct 20Oct 21Oct 22Oct 242026-10-20: n 9 → 10; joined: together2026-10-19 UTC: 0.1 USD2026-10-20 UTC: 0.1 USD2026-10-21 UTC: 0.1 USD2026-10-22 UTC: 0.1 USD2026-10-23 UTC: 0.1 USD2026-10-24 UTC: 0.1 USD

Change: 0.0% from 2026-10-19 to 2026-10-24 (like for like, on the constituents both days share).

Dashed lines with a diamond mark days on which the set of constituents (n) changed; a jump on such a day can be coverage, not a price move.

By host

8 lines, 2026-10-19 to 2026-10-20; values between 0.1 and 1.04 USD; latest 1.04 USD.0.02480.29740.570.84261.12Oct 19Oct 20Oct 21Oct 22Oct 23SambaNova: 2026-10-20 UTC: 0.6 USDTogether: 2026-10-20 UTC: 1.04 USDDeepInfra: 2026-10-19 UTC: 0.1 USDNovita: 2026-10-19 UTC: 0.135 USDAkashML: 2026-10-19 UTC: 0.2 USDParasail: 2026-10-19 UTC: 0.22 USDCloudflare: 2026-10-19 UTC: 0.293 USDSambaNova: 2026-10-19 UTC: 0.45 USD

One line per host, from that listing's own recorded changes; a price holds until the next change.

Hosts with their own price page (first-party)

Hosts publish no quantization on their pages, so it reads unknown.

SKUHostQuantizationInput / 1MOutput / 1MCached inputContext windowLast verifiedSource
inference/sambanova/meta-llama/llama-3.3-70b-instruct/sambanova/unknownSambaNovaunknownUSD 0.60USD 1.20131,0722026-10-23 18:46 UTCsambanova-models
inference/together/meta-llama/llama-3.3-70b-instruct/together/unknownTogetherunknownUSD 1.04USD 1.042026-10-23 01:10 UTCtogether-pricing

Hosts via OpenRouter endpoints

The upstream host, its quantization and the price OpenRouter lists for it (a cross-check of the host's own price).

SKUHostQuantizationInput / 1MOutput / 1MCached inputContext windowLast verifiedSource
inference/openrouter/meta-llama/llama-3.3-70b-instruct/deepinfra/fp8DeepInfrafp8USD 0.10USD 0.32131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/novita/bf16Novitabf16USD 0.135USD 0.4012,2882026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/akashml/fp8AkashMLfp8USD 0.20USD 0.52USD 0.10131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/parasail/fp8Parasailfp8USD 0.22USD 0.50USD 0.11131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/cloudflare/fp8Cloudflarefp8USD 0.293USD 2.25324,0002026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/sambanova/unknownSambaNovaunknownUSD 0.45USD 0.90131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/groq/unknownGroqunknownUSD 0.59USD 0.79USD 0.295131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/coreweave/fp16CoreWeavefp16USD 0.71USD 0.71USD 0.71128,0002026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/google/unknownGoogleunknownUSD 0.72USD 0.72128,0002026-10-23 22:54 UTCopenrouter-endpoints community cross-check
inference/openrouter/meta-llama/llama-3.3-70b-instruct/together/unknownTogetherunknownUSD 1.04USD 1.04131,0722026-10-23 22:54 UTCopenrouter-endpoints community cross-check

Other community cross-checks

Not in the cheapest-host series.

SKUHostQuantizationInput / 1MOutput / 1MCached inputContext windowLast verifiedSource
inference/hf-router/meta-llama/llama-3.3-70b-instruct/novita/unknownNovitaunknownUSD 0.135USD 0.4012,2882026-10-23 18:46 UTChf-inference-providers community cross-check
inference/hf-router/meta-llama/llama-3.3-70b-instruct/ovhcloud/unknownovhcloudunknownUSD 0.74USD 0.74131,0722026-10-23 18:46 UTChf-inference-providers community cross-check
inference/hf-router/meta-llama/llama-3.3-70b-instruct/scaleway/unknownscalewayunknownUSD 1.026USD 1.0262026-10-23 18:46 UTChf-inference-providers community cross-check

How this page is built

It reads the latest value of the Cheapest Llama 3.3 70B Instruct input price index series and the per-provider detail behind it, so the page and the series always agree. Definition: The lowest input price, in USD per million tokens, for Llama 3.3 70B Instruct among the hosts listing it (OpenRouter endpoints and first-party host price pages), with the host and quantization of the cheapest endpoint stated. Data: JSON · CSV · Atom feed.

Cite this

Free to reuse under CC BY 4.0 with this credit. The numbers are list prices, not transacted prices.

InfraRabbit (2026). Llama 3.3 70B Instruct API price [data as of 2026-10-24]. https://infrarabbit.com/llm/llama-3-3-70b-price (accessed 2026-10-24). List prices, not transacted prices. CC BY 4.0.
BibTeX
@misc{infrarabbit_llm_llama_3_3_70b_price_2026,
  author       = {{InfraRabbit}},
  title        = {Llama 3.3 70B Instruct API price},
  year         = {2026},
  howpublished = {\url{https://infrarabbit.com/llm/llama-3-3-70b-price}},
  note         = {Data as of 2026-10-24; accessed 2026-10-24; list prices, not transacted prices; CC BY 4.0}
}

Related