Skip to content
ModelCap

Llama 3.3 70B Instruct

Current board #70Arena #234Gated access

Meta·meta-llama/llama-3.3-70b-instruct

Llama 3.3 70B Instruct is a language model published by Meta. 13 OpenRouter endpoints are published in the latest snapshot. It accepts text input and returns text. The published context limit is 131K tokens. The catalogue declares support for tool use, structured outputs.

ModelCap Index · Current board
41.4#70
Measured · 85% support · score interval 41.4–47.8
Market Gravity
44.0
Secondary market signal
Input
$0.10
per 1M tokens
Output
$0.32
per 1M tokens
Context
131K
tokens
Providers
13
endpoints
Gateway spend
—
latest · Vercel share

Providers

Uptime measured over the last 30 minutes
ProviderInputOutputContextQuantUptime
DeepInfraturboCheapest$0.10$0.32131Kfp897.30%
Nebiusfp8$0.13$0.40131Kfp8—
AkashMLfp8$0.13$0.40131Kfp898.65%
Novitabf16$0.135$0.4012Kbf1697.63%
Parasailfp8$0.22$0.50131Kfp8—
CoreWeavefp16$0.71$0.71128Kfp16100.0%
Google$0.72$0.72128K—97.41%
Googleus-central1$0.72$0.72128K—100.0%
Crusoebf16$0.25$0.75131Kbf16100.0%
Groq$0.59$0.79131K—100.0%
SambaNova$0.45$0.90131Kbf1694.73%
Togetherfp8$1.04$1.04131Kfp899.68%
Cloudflarefp8$0.293$2.2524Kfp899.71%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

ProviderInputOutputContextTTFTThroughputFeatures
novita
live
$0.135$0.4012K419 ms37.9 tok/stools
ovhcloud
live
$0.74$0.74131K382 ms27.0 tok/stools · structured
groq
live
$0.59$0.79131K394 ms257.4 tok/stools
scaleway
live
$1.03$1.03—402 ms81.3 tok/stools · structured
together
live
$1.04$1.04131K1171 ms51.5 tok/stools · structured
featherless-ai
live
——————

Reported evaluation leads (11)

Excluded from score

Structured rows reported in this exact model repository at the pinned revision below. They are useful research leads, but remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.

Dataset / taskMetricValueRevisionProvenance
Idavidrein/gpqa
diamond
—51.5152
Unpinned
not published
Reported
Quarantined
LEXam-Benchmark/LEXam
open_question
—41.27
Unpinned
not published
Reported
Quarantined
LEXam-Benchmark/LEXam
mcq_4_choices
—28.19
Unpinned
not published
Reported
Quarantined
openai/gsm8k
gsm8k
—94.8446
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
src
—0.4167
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
time
—0.75
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
access_deny
—0.0833
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
memory_label
—0.5833
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
injection
—0.5833
Unpinned
not published
Reported
Quarantined
thamilvendhan/signalbench
bot_policy
—0.0833
Unpinned
not published
Reported
Quarantined
TIGER-Lab/MMLU-Pro
mmlu_pro
—65.92
Unpinned
not published
Reported
Quarantined

Other versions (1)

ModelCap Index · Current board

Methodology
41.4
Measured · 3 public benchmark observations across 3 boards
Current board
#70
Public rank basis
Measured (measured)
Index support
85.3% · strong
Index interval
41.4–47.8
Rank posterior
Not available
Top 5 / Top 10 probability
Not available / Not available
Posterior as of
Snapshot timestamp unavailable
Identity binding
Exact catalogue product · meta-llama/meta-llama/llama-3.3-70b-instruct · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata meta-llama/Llama-3.3-70B-Instruct@6f6073b423013f6a7d4d9f39144961bfbfbc386b (not the evaluation revision)
Direct benchmark observations
3
Breadth-adjusted BES point estimate
44.6
Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability
44.6
Evidence status
confirmed · 94% mass
General preference
45.3
Coding
42.4
Agents & tools
41.2
Reasoning
—
Evidence breadth
97%
Evidence coverage
85%
Market and catalogue signals · secondary, never the language Index rank
Usage (OpenRouter popularity)
20.8
Liquidity (providers × uptime)
98.1
Open reach (HF downloads)
—
Surface (context / tools / modalities)
45.2
Freshness
9.8
  • · OpenRouter popularity #91
  • · 12 live providers
  • · Catalogue: context 131072, tools
  • · First seen 2024-12-06

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

BES evidence selected · measured input only

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Deployment readiness

Metadata index
88.2
strong · 93% metadata coverage
Not capability
Artifact reproducibility
100.0
Access & legal clarity
68.3
Deployability
97.9
Evaluation provenance
75.7

Missing: accessLegalClarity.resolvable-license-terms, evaluationProvenance.evaluation-dataset-revisions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage
Market Gravity
44.0
OpenRouter weekly popularity
#91 of 298

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

Status
Ranked
Source
Arena
Official rank
#234 of 385
Rating
1318.3
95% confidence interval
1315–1322
Votes
55K
Source snapshot
6 Aug 2026, 13:00 UTC
Category
overall
Identity match confidence
100%

Market Gravity breakdown

Usage55%
20.8

OpenRouter weekly popularity across the full model catalogue.

Liquidity25%
98.1

Independent providers versus the model's open or closed cohort, weighted by uptime.

Open reach15%
50.0

Hugging Face 30-day downloads for open models; neutral for closed models.

Freshness5%
9.8

Time since first public availability, on a six-month half-life.

Specification

Context window
131K tokens
Max output
16K tokens
Inputs
Text
Outputs
Text
Tokenizer
Llama3
Cached input tokens
Not offered
First seen on OpenRouter
6 Dec 2024
Knowledge cutoff
2023-12-31

Capabilities

  • Not supported: Reasoning
  • Supported: Tool use
  • Supported: Structured output
  • Supported: Response format
  • Not supported: Moderated

Weights & access

Gated access. The weights are hosted behind an access request or acceptance step. Review the publisher's access requirements and license before use or redistribution.

Access
Gated access
Licence
llama3.3
Pinned revision
6f6073b42301
Parameters
70.6B
Architecture
LlamaForCausalLM
Model type
llama
Hugging Face downloads (30d)
394K
Downloads (all time)
13.5M
Likes
3K
Repository
meta-llama/Llama-3.3-70B-Instruct
Repository updated
21 Dec 2024, 18:28 UTC
Base lineage
meta-llama/Llama-3.1-70B