Llama 3.3 70B Instruct
Meta · released Nov 26, 2024 · meta-llama/Llama-3.3-70B-Instruct
- Type
- Open weightsllama3.3
- Params
- 70.6B
- Context
- 131K
about 98K words of context · download allowed, licence restricts use
Our take
Written Sep 2, 2026Llama 3.3 is a 70.6-billion-parameter text model from Meta with a 131,072-token request limit and a wide spread of hosted providers. Released in late 2024, it is a solid workhorse for instruction-heavy pipelines and long documents, though its science reasoning is weak.
Pick this for long-document work at 131K tokens when you need open-weights flexibility, or for budget inference on the cheapest hosts. Use it for instruction-heavy pipelines where its 90% accuracy on that benchmark matters. Skip it if you need graduate-level science reasoning, or if you want the cheapest host but need the throughput only a premium GPU cloud delivers.
The case for it
- Strong measured instruction-following accuracy at 90% on IFEval.
- Broad provider choice with a sevenfold price spread between budget and premium hosts.
- 131,072-token request limit for the full 70.6-billion-parameter model.
- Coding preference score leads its own measured sub-scores by a 28-point gap.
The case against it
- Graduate-level science reasoning is a clear gap at 10.5% on GPQA Diamond.
- General knowledge breadth trails on harder MMLU at 48.1% correct.
- Premium hosts charge several times the budget rate with no measured quality gain.
How good is it?
An open text model for chat and general assistance, though it trails most models on everyday questions, writing and coding.
- getting answers to everyday questionsArena Text (overall) · 138th of 168
- drafts, rewrites and editingArena Creative Writing · 136th of 168
- writing and completing codeArena Coding · 146th of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)138th of 168 · 1318
CodingWriting and fixing code on its own
Arena Coding146th of 168 · 1346
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing136th of 168 · 1285
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model9 scoresEvery figure we hold, from 9 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Max (32-core GPU) · 64 GB
Borderline fit on an estimated size. It leaves 0.6 GB spare on a size we calculated rather than measured, and a 10% error either way would change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 4 hours ago — each listing carries its own date.
- per 1M tokens
- $0.10 in / $0.32 out
- Context served
- 131K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| DeepInfraturbo tierfp8Through OpenRouter | $0.10 / $0.32checked 4 hours ago | 131K16K max reply | 13 tok/s | No | No | Unknown |
| OpenRouterOpenRouter's own listing | $0.10 / $0.32checked 4 hours ago | 131K | not measured | Unknown | Unknown | Unknown |
| Novita AIbf16Direct and through OpenRouter | $0.14 / $0.40checked 4 hours ago | 12K11K max reply through OpenRouter | 8 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| Parasailfp8Through OpenRouter | $0.22 / $0.50checked 4 hours ago | 131K16K max reply | 20 tok/s | No | No | Confirmed |
| AkashMLfp8Through OpenRouter | $0.20 / $0.52checked 4 hours ago | 131K128K max reply | 20 tok/s | No | No | Confirmed |
| CoreWeavefp16Through OpenRouter | $0.71 / $0.71checked 4 hours ago | 128K115K max reply | 51 tok/s | No | No | Confirmed |
| Google Vertex AIThrough OpenRouter | $0.72 / $0.72checked 4 hours ago | 128K115K max reply | 42 tok/s | No | No | Confirmed |
| Google Vertex AIus-central1Through OpenRouter | $0.72 / $0.72checked 4 hours ago | 128K8K max reply | 81 tok/s | No | No | Confirmed |
| GroqThrough OpenRouter | $0.59 / $0.79checked 4 hours ago | 131K33K max reply | 156 tok/s | No | No | Confirmed |
| SambaNovaThrough OpenRouter | $0.45 / $0.90checked 4 hours ago | 131K3K max reply | 114 tok/s | Unknown | Unknown | Confirmed |
| Together AIThrough OpenRouter | $1.04 / $1.04checked 4 hours ago | 131K2K max reply | 9 tok/s | No | No | Confirmed |
| SambaNovaDirect | $0.60 / $1.20checked 4 hours ago | 131K | not measured | Unknown | Unknown | Unknown |
| Cloudflare Workers AIfp8Through OpenRouter | $0.29 / $2.25checked 4 hours ago | 24K22K max reply | 11 tok/s | No | Yesunknown period | Unknown |
DeepInfra
Novita AI
Google Vertex AI
Groq
SambaNova
Together AI
SambaNova
Cloudflare Workers AIAcross the 13 listings we hold: 10 say they do not train on prompts (1 of them only through OpenRouter), 0 say they do and 3 do not say. 9 appear in the zero-retention registry we check (1 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| DeepInfraturbo · fp8Through OpenRouter | ✓ | ✓ | ✓ |
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Novita AIbf16Direct and through OpenRouter | ✓ | ✗ | ✗ |
| Parasailfp8Through OpenRouter | ✗ | ✓ | ✓ |
| AkashMLfp8Through OpenRouter | ✓ | ✓ | ✓ |
| CoreWeavefp16Through OpenRouter | ✓ | ✓ | ✓ |
| Google Vertex AIThrough OpenRouter | ✗ | ✓ | ✗ |
| Google Vertex AIus-central1Through OpenRouter | ✓ | ✓ | ✓ |
| GroqThrough OpenRouter | ✓ | ✓ | ✗ |
| SambaNovaThrough OpenRouter | ✗ | ✓ | ✓ |
| Together AIThrough OpenRouter | ✓ | ✓ | ✓ |
| SambaNovaDirect | |||
| Cloudflare Workers AIfp8Through OpenRouter | ✗ | ✓ | ✗ |
Tool calling: 8 of 13 listings say yes, 4 say no, 1 publishes no parameter list. JSON output: 11 of 13 listings say yes, 1 says no, 1 publishes no parameter list. Strict schema: 8 of 13 listings say yes, 4 say no, 1 publishes no parameter list.
Models people weigh against Llama 3.3 70B Instruct
When we formed this view
Recent changes
What moved
input +54% ($0.13 → $0.20 per 1M tokens), output +30% ($0.40 → $0.52 per 1M tokens)What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 1 of 13 listings publishes no parameter list, so what its API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 3 of 13 listings do not say whether they train on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsllama3.3, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
llama3.3
License tag "llama3.3" imported from Hugging Face; terms pending curation — review the original text before relying on it.
Identifiers
- Hugging Face
- meta-llama/Llama-3.3-70B-Instruct
- Takes in, gives back
- Text in, text out
- Catalogue slug
- meta-llama-llama-3-3-70b-instruct