Llama 2 70B Chat Hf
Jul 2023
Llama 2 70B Chat HfCompare inon
Llama 2 70B Chat Hf measured on 3 of 104 Chat & Knowledge benchmarks3
BenchmarkByScoreRank
Legacy · saturated or no longer reported2
No benchmark matches
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperGPT-5.6 Lunamax×3.5-34%$0.18
OpenQwen3.6 27Bunstated×3.3+99%$0.60 / $3.60
CheaperMiniMax M3unstatedOpen×3.5-23%$0.30 / $1.20Memory (4-bit): 41 GB → 256 GB+1CheaperGPT-6 Lunamax×3.4-71%$0.068Function calling only with none · Chat Completions+3
CheaperGLM 5.3 FlashmaxOpen×3.4-65%$0.28Memory (4-bit): 41 GB → 193 GB+5ModelScoreCostPer taskWhat changes
GPT-5.6 Lunamax×3.5-34%$0.18MiniMax M3unstatedOpen×3.5-23%$0.30 / $1.20Memory (4-bit): 41 GB → 256 GB+1GPT-6 Lunamax×3.4-71%$0.068Function calling only with none · Chat Completions+3GLM 5.3 FlashmaxOpen×3.4-65%$0.28Memory (4-bit): 41 GB → 193 GB+5
Claude Haiku 5.5max×3.4-71%$0.21Remove sampling parameters+5ModelScoreCostPer taskWhat changes
GPT-5.6 Lunalow×3.1-34%$0.01GPT OSS 20BhighOpen+75%-95%$0.011
Gemini 3.1 Flash Litehigh×3.1-17%$0.06GPT-6 Lunamax×3.4-71%$0.068Function calling only with none · Chat Completions+3GPT OSS 120BhighOpen×2.9-91%$0.092Memory (4-bit): 41 GB → 70 GBModelScoreCostPer taskWhat changes
Qwen3.6 27BunstatedOpen×3.3+99%$0.60 / $3.60Qwen3.6 35B A3BnoneOpen×3.2-18%$0.25 / $1.49Qwen 3.5 35B A3BunstatedOpen×3.2same$0.25 / $2Qwen 3.5 9BunstatedOpen×3-86%$0.08 / $0.13Gemma 4 31BminimalOpen×2.9-76%$0.10 / $0.34ModelScoreCostPer taskWhat changes
Muse Sparkunstated×3.4—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Muse Spark Open 816B A42B×3.4—$1.25 / $4.25Context: not verifiedLlama 4 Maverick 17B 128E Instruct FP8Open×2.5—$0.20 / $0.80Memory (4-bit): 41 GB → 241 GBLlama 4 ScoutunstatedOpen+97%—$0.10 / $0.30Memory (4-bit): 41 GB → 65 GBLlama 3.1 405B Instruct+93%——Llama 3.3 70B InstructunstatedOpen+80%—$0.10 / $0.32Memory (4-bit): 41 GB → 42 GBLlama 3.1 70B InstructunstatedOpen+68%—$0.40 / $0.40Memory (4-bit): 41 GB → 42 GBLlama 3.2 90B Vision InstructOpen+56%—$0.35 / $0.40Memory (4-bit): 41 GB → 53 GBAll Meta models →Run it yourself
estimateFits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✓ 48 GB GPU · compressed✓ 80 GB GPU · 8-bit
LicenceLlama 2 licencecommercial use with conditions
FAQ
How much does Llama 2 70B Chat Hf cost?›
$0.64 per 1M input tokens and $0.80 per 1M output tokens (cheapest host: DeepInfra).
Is Llama 2 70B Chat Hf free to use?›
There is no free API; the cheapest paid price is $0.64 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Llama 2 70B Chat Hf, and what is it best at?›
#125 of 133 in Chat & Knowledge.
Is Llama 2 70B Chat Hf better than Llama 3.1 8B Instruct?›
On GPQA Diamond · Epoch, Llama 2 70B Chat Hf scores 26.3 vs 27 — about the same.
How can I use Llama 2 70B Chat Hf?›
Through the API of DeepInfra.
What is the context window of Llama 2 70B Chat Hf?›
4K tokens.
When was Llama 2 70B Chat Hf released? Is there a newer version?›
Released 2023-07-14. It is the current version in its line.
Is Llama 2 70B Chat Hf open source?›
Yes, the weights are open (llama2 licence).
Can I run Llama 2 70B Chat Hf locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Llama 2 70B Chat Hf read images?›
No, it accepts text only.
What is Llama 2 70B Chat Hf?›
Llama 2 70B Chat Hf is an open-weights language model from Meta, released Jul 2023.
What are the best alternatives to Llama 2 70B Chat Hf?›
GPT-5.6 Luna (cheaper), Qwen3.6 27B (open weights).