LLM Rank.top

Leaderboard · Model · Updated

Llama 3.3 70B Instruct

Meta's 70B open-weights model — frontier-class for its size, runs on prosumer GPUs.

Composite 64.7 Rank #24 of 30 Meta open-weights Open-weights (Llama 3.3 Community License) Released 2024-12
Try Llama 3.3 70B Instruct → Compare with Qwen2.5 72B Instruct → Or route via OpenRouter →

At a glance

ProviderMeta
Released2024-12
Tieropen-weights
LicenseOpen-weights (Llama 3.3 Community License)
Modalitiestext
Context window128k tokens
Max output4.096k tokens
API price · input$0.23 / 1M tokens
API price · output$0.40 / 1M tokens
Hugging Facemeta-llama/Llama-3.3-70B-Instruct

Benchmark performance

How Llama 3.3 70B Instruct stacks up against the current leader (GPT-5) and the median model in the leaderboard:

BenchmarkLlama 3.3 70B InstructGPT-5Median
Chatbot Arena Elo 1257 1410 1320
MMLU-Pro 68.9 86.8 78.0
GPQA Diamond 50.5 87.3 65.0
MATH 77.0 96.7 78.3
HumanEval 88.4 95.1 92.0
SWE-Bench Verified N/A 74.9 49.0

Numbers compiled from provider technical reports and Chatbot Arena. See methodology for the composite-score formula.

Don't lock in to one provider.

OpenRouter routes your requests across Llama 3.3 70B Instruct, GPT-5, Claude, Gemini, and 100+ other models behind a single API key — pay-as-you-go, no monthly minimum. Try OpenRouter → (affiliate · supports this site)

What does Llama 3.3 70B Instruct cost in practice?

API pricing is $0.23 per 1M input tokens and $0.40 per 1M output tokens. Assuming a 50/50 input/output split, here is what that looks like at three workload sizes:

VolumePer dayPer monthPer year
1M tokens/day$0.32$9.45$114.97
10M tokens/day$3.15$94.50$1,150
100M tokens/day$31.50$945.00$11,498

Strengths & weaknesses

Where it shines

No clear top-tier strengths — this is a mid-pack model.

Where it lags

  • Arena: 1257 (rank #23 of 27, below average)
  • MMLU-Pro: 68.9 (rank #24 of 29, below average)
  • GPQA: 50.5 (rank #22 of 29, mid-pack)

Best alternatives

The closest models to Llama 3.3 70B Instruct by tier and benchmark score:

ModelScore$ in / outContextAction
Qwen2.5 72B Instruct
Alibaba
65.6 $0.35 / $0.40 131.072k Try → · vs Llama
Llama 3.1 405B Instruct
Meta
65.7 $2.70 / $2.70 128k Try → · vs Llama
DeepSeek V3
DeepSeek
68.0 $0.27 / $1.10 128k Try → · vs Llama
Qwen2.5-Coder 32B
Alibaba
68.8 $0.18 / $0.18 131.072k Try → · vs Llama
Llama 3.1 70B Instruct
Meta
60.2 $0.23 / $0.40 128k Try → · vs Llama

Frequently asked questions

Is Llama 3.3 70B Instruct a good model?

Llama 3.3 70B Instruct scores 64.7 on the llmrank.top composite (rank #24 of 30). It trails the frontier — for top performance, look at GPT-5 (86.0). Whether it's the right fit depends on your workload — see the use-case discussion on this page.

How much does Llama 3.3 70B Instruct cost?

Llama 3.3 70B Instruct is priced at $0.23 per 1M input tokens and $0.40 per 1M output tokens (USD). For a 10M-token-per-day workload split 50/50, that works out to roughly $3.15/day or $1,150/year.

What is Llama 3.3 70B Instruct's context window?

128k tokens. For very long inputs, consider a 1M+ context model like Gemini 2.5 Pro or GPT-4.1.

Is Llama 3.3 70B Instruct open source?

Yes — released under the Llama 3.3 Community License license, weights are downloadable from the provider or Hugging Face.

What is Llama 3.3 70B Instruct's SWE-Bench score?

Llama 3.3 70B Instruct scores not reported on SWE-Bench Verified — the benchmark that measures real-world GitHub issue resolution. Its Chatbot Arena Elo is 1257.

What are the best alternatives to Llama 3.3 70B Instruct?

Closest alternatives by tier and score: Qwen2.5 72B Instruct, Llama 3.1 405B Instruct, DeepSeek V3. See the alternatives section on this page for side-by-side numbers.


Related: Qwen2.5 72B Instruct · Llama 3.1 405B Instruct · DeepSeek V3 · Full leaderboard

Spotted out-of-date numbers? Open an issue — corrections usually ship within 24h.

Try Llama 3.3 70B Instruct now

Direct link to the official playground — or use OpenRouter to A/B test it against any other model on a single API key.

Open Meta playground → Try via OpenRouter →