Get started

Llama 3.3 70B Instruct

Sold by 7 sellers, timed at 76 tok/s, scored 11.9 on the Coding Index, and with results on 1 openly licensed board.

retrieved sha256 · 0bb53a77713f

Models/Meta Llama/Llama 3.3 70B InstructUSD per 1M tokens · as OpenRouter routes it
meta-llama/llama-3.3-70b-instructMeta Llama · base listing · Llama3 tokenizer
OpenRouter list price · the routed rate, as the catalog carried itretrieved 17 Nov 16:29 UTC from openrouter.ai ↗
Input$0.71per 1M tokens
Output$0.71per 1M tokens
Cached input$0.71per 1M tokens
Context131k131,072 tokens
What Vitna read · 5 providers, each from its own pricing page, kept as dated recordsin / out per 1M tokens
Googleopen weights$0.72 / $0.72confirmed 22 Nov 2026
AWSopen weights$0.72 / $0.72confirmed 22 Nov 2026
DeepInfraopen weights$0.10 / $0.32confirmed 22 Nov 2026
Parasailopen weights$0.22 / $0.50confirmed 22 Nov 2026
Novita AIopen weights$0.135 / $0.40confirmed 22 Nov 2026

The same weights served by 5 providers, each at its own list price. OpenRouter's rate is its routed rate across whichever of them it uses, and it need not match any one of them.

What the list says · the fields OpenRouter publishes for this id9 fields
Listed
Context131,072 tokens
Max output115,200 tokens
Modalitiestext → text
TokenizerLlama3
Moderationnot moderated by OpenRouter
Knowledge cutoff2023-12-31
What the weights say · read from the model's own repository on Hugging Faceretrieved from huggingface.co ↗
Licencellama3.3
Parameters70.6B · 70,553,706,496 declared
Model classLlamaForCausalLM
Downloadgated, each request approved by the maker
Published

What sellers charge

7 sellers · 7.2x from cheapest to dearest

One tick per seller, on a log scale of input price per million from $0.10 at DeepInfra to $0.72 at Amazon Bedrock. 5 sellers undercut the $0.71 listed; 1 is dearer. A faint tick declares no number format. Every seller is named on the endpoint ledger below.

The same sellers grouped by the number format each declares, on the same scale. The fp8 group alone spans 54% of the range, so the cheap end is not cheap for its number format. A declaration is checked by nobody: this groups the rows and certifies none.

Measured speed

45,443 requests · 7 endpoints · 30-minute window
Throughput76 tok/sMedian output rate, weighted across the 7 endpoints serving it by the requests each answered.the speed board
First token632msMedian wait before the first token, over the same window and weighted the same way.how Vitna reads a figure
Endpoint spread18 to 220The slowest and the fastest route serving this model, in tokens per second. Speed is a property of the route, not of the weights.the board, fastest first

OpenRouter's own timing of its routed traffic, weighted by requests, over one 30-minute window read . Every route is on the endpoint ledger below.

Every endpoint

7 priced · 7 timed · cheapest first
  1. DeepInfradeepinfra/turbo$0.10$0.32not listedfp816k18 tok/s9856ms5,055 req99.39%
  2. AkashMLakashml/fp8$0.20$0.52$0.10fp8128k30 tok/s24769ms9,658 req97.91%
  3. Parasailparasail/fp8$0.22$0.50$0.11fp816k28 tok/s111.17s8,737 req100%
  4. Cloudflarecloudflare/fp8$0.293$2.253not listedfp822k40 tok/s18.1965ms1,184 req98.80%
  5. Groqgroq$0.59$0.79$0.295not declared33k220 tok/s149490ms7,352 req99.72%
  6. CoreWeavecoreweave/fp16 · listed$0.71$0.71$0.71fp16115k86 tok/s73143ms13,102 req100%
  7. Amazon Bedrockamazon-bedrock/us-east-1$0.72$0.72not listednot declared8k98 tok/s80346ms355 req100%

Each track runs from a route's p90, the rate nine in ten of its requests met or beat, up to its median. Weighted by requests, the figure above reports where the traffic went rather than what one route does: 4 of the 7 timed routes are slower than it, and the middle route runs at 40 tok/s. The widest median-to-p90 gap is Parasail's, 2.55x.

Uptime is OpenRouter's own figure for each seller over the trailing day, read once a day, and the column prints the latest reading, taken . 8 of the 10 sellers with a reading fell under 99% on at least one of the 23 days held, 23 Oct 2026 to 21 Nov 2026: SambaNova at 88.45% on 24 Oct 2026, CoreWeave at 93.52% on 16 Nov 2026, Novita at 95.72% on 25 Oct 2026, Cloudflare at 96.03% on 14 Nov 2026, and 4 more. Source: OpenRouter's public endpoints API, one reading a day for each seller: 1,211 sellers carried one on the latest of the 23 days held.

Source: OpenRouter's public endpoints API. 1257 sellers across 359 models when the mirror ran, 186 of them served by more than one. List prices, not what anybody paid; a discount is the seller's own. Read . Speed is OpenRouter's own timing over one 30-minute window, read .

What each seller accepts

17 parameters · 7 sellers
  1. max_tokensevery seller
  2. stopevery seller
  3. temperatureevery seller
  4. top_pevery seller
  5. response_format6 of 7not Amazon Bedrock
  6. seed6 of 7not Amazon Bedrock
  7. top_k6 of 7not Groq
  8. frequency_penalty5 of 7not Amazon Bedrock and Groq
  9. presence_penalty5 of 7not Amazon Bedrock and Groq
  10. repetition_penalty5 of 7not Amazon Bedrock and Groq
  11. tool_choice5 of 7not Cloudflare and Parasail
  12. tools5 of 7not Cloudflare and Parasail
  13. logprobs4 of 7not Amazon Bedrock, DeepInfra, and Groq
  14. structured_outputs4 of 7not Amazon Bedrock, Cloudflare, and Groq
  15. top_logprobs4 of 7not Amazon Bedrock, DeepInfra, and Groq
  16. logit_bias3 of 7only Cloudflare, DeepInfra, and Parasail
  17. min_p2 of 7only Cloudflare and DeepInfra

4 of these 17 are accepted by every one of the 7 sellers that publish a list; the other 13 are declared by some sellers and not others, so what a request sending one of them gets depends on which seller serves it. Each list is the seller's own declaration in OpenRouter's endpoints API, read , and nothing here tested one.

Price position

342 priced models

Blended $0.71 per million tokens, 3 in to 1 out: cheaper than 182 of the 342 priced models in the catalog. Each bar is one stretch of the catalog on a logarithmic price scale; the mark is this listing.

Measured quality

145 models with an index · 132 on a design arena board
Intelligence Indexnot scoredArtificial Analysis has not scored this model on its Intelligence Index.the intelligence leaderboard
Coding Index11.9Position 109 of 118, above 8% of the models scored.the coding leaderboard
Agentic Indexnot scoredArtificial Analysis has not scored this model on its Agentic Index.the agentic leaderboard

Not ranked in any Design Arena category.

Epoch Capabilities Index · a fit by a third party over every benchmark this model has results onthe whole index, with intervals

Epoch AI has not placed Llama 3.3 70B Instruct on its Capabilities Index, or places it under a name the pairing does not reach. Epoch places a model once four of its benchmarks carry a result.

Benchmark results

1 of 6 boards · 3 results
Epoch AI Benchmarking Hub · publisher-run, CC-BY-4.03 of 22 tasks · effort unstated
TaskAll runsScore
GPQA diamond3420.47
MATH level 51080.42
OTIS Mock AIME 2024-20253190.05

Mean accuracy, 0 to 1, on the task the row names, as Epoch AI runs it and averaged over repeated runs. All eighteen of their tasks, including the held-out FrontierMath tiers, named as they name them. The model id carries the effort setting the run used. Epoch AI Benchmarking Hub names this run Llama-3.3-70B-Instruct, with no effort setting written on it. All runs is how many runs the publisher scored on the task, every model's. Read from epoch.ai ↗, which dates its own data . Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.

1 of the 6 openly licensed boards Vitna mirrors carries a result for Llama 3.3 70B Instruct, each printed as its publisher published it and none added to another. MTEB, LiveBench, HELM, Vectara hallucination leaderboard and Aider polyglot have not published one under a name that matches its listing, and names are matched to the catalog exactly and never guessed. A task that is a link opens its column on /capability, with every run scored on it; an effort that is a link opens its run's row there, where the run is ranked on Vitna's scale. Every board, with its licence and what its publisher measures, is on /benchmarks.

Sources

Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Source: each model's own Hugging Face repository (huggingface.co), as of 20 Nov 2026. 162 repositories read, 4 did not answer.
Licence, parameter count, model class and gating: each model's own Hugging Face repository, the one OpenRouter names for the listing.
A licence is the tag the uploader put on the repository, read from nothing else and verified by nobody.
A parameter count is the total the repository's safetensors files declare; for a mixture of experts that is every expert, not the share one token runs through.
Vitna ran none of these evaluations and adjusted none of these figures.