DeepSeek R1 Distill Llama 70B
Jan 2025
DeepSeek R1 Distill Llama 70BCompare inon
DeepSeek R1 Distill Llama 70B 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+64%-38%$0.18
OpenQwen3.6 27Bunstated+54%+86%$0.60 / $3.60
CheaperMiniMax M3unstatedOpen+63%-28%$0.30 / $1.20Memory (4-bit): 42 GB → 256 GB+2CheaperGPT-6 Lunamax+62%-72%$0.068Function calling only with none · Chat Completions+3
CheaperGLM 5.3 FlashmaxOpen+62%-67%$0.28Memory (4-bit): 42 GB → 193 GB+5ModelScoreCostPer taskWhat changes
GPT-5.6 Lunamax+64%-38%$0.18MiniMax M3unstatedOpen+63%-28%$0.30 / $1.20Memory (4-bit): 42 GB → 256 GB+2GPT-6 Lunamax+62%-72%$0.068Function calling only with none · Chat Completions+3GLM 5.3 FlashmaxOpen+62%-67%$0.28Memory (4-bit): 42 GB → 193 GB+5
Claude Haiku 5.5max+61%-72%$0.21Remove sampling parameters+5ModelScoreCostPer taskWhat changes
GPT-5.6 Lunalow+48%-38%$0.01
Gemini 3.1 Flash Litehigh+47%-22%$0.06GPT-6 Lunamax+62%-72%$0.068Function calling only with none · Chat Completions+3GPT OSS 120BhighOpen+36%-92%$0.092Memory (4-bit): 42 GB → 70 GBGPT-5.6 Terralow+57%+521%$0.14Prompts with >272K input tokensModelScoreCostPer taskWhat changes
Qwen3.6 27BunstatedOpen+54%+86%$0.60 / $3.60Qwen3.6 35B A3BnoneOpen+52%-23%$0.25 / $1.49Qwen 3.5 35B A3BunstatedOpen+50%-5%$0.25 / $2Qwen 3.5 9BunstatedOpen+42%-87%$0.08 / $0.13Gemma 4 31BminimalOpen+36%-78%$0.10 / $0.34ModelScoreCostPer taskWhat changes
DeepSeek V4 Pro 0813maxOpen+64%—$0.67Memory (4-bit): 42 GB → 990 GB+5R1 0528unstatedOpen+37%—$0.39 / $1.63Context: 131K → 128K+1DeepSeek V3unstatedOpen+28%—$0.28 / $0.42Context: 131K → 128K+1DeepSeek V3 0324unstatedOpen+21%—$0.24 / $0.90Context: 131K → 128K+1DeepSeek R1 Distill Qwen 32BOpen+15%—$0.27 / $0.27No strict JSONDeepSeek R1 0528 Qwen3 8BunstatedOpen-83%—$0.06 / $0.09Context: 131K → 33KDeepSeek V4.1 Flash 1017maxOpenn/a—$0.26Memory (4-bit): 42 GB → 458 GB+5DeepSeek V3.2 SpecialethinkingOpenn/a——Context: 131K → 128K+1All DeepSeek models →Run it yourself
estimateFits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✓ 48 GB GPU · compressed✓ 80 GB GPU · 8-bit
LicenceMITfree for commercial use
Benchmarks
in Chat & KnowledgeGPQA Diamond · Epoch55.7#98
MATH Level 589.9#6
OTIS Mock AIME 2024 202551.4#92
AIME 2024 pass@1lab70
AIME 2024 cons@64lab86.7
MATH-500 pass@1lab94.5
GPQA Diamond pass@1lab65.2
FAQ
How much does DeepSeek R1 Distill Llama 70B cost?›
$0.70 per 1M input tokens and $0.80 per 1M output tokens (cheapest host: DeepInfra).
Is DeepSeek R1 Distill Llama 70B free to use?›
There is no free API; the cheapest paid price is $0.70 per 1M input tokens. The weights are open, so you can run it yourself.
How good is DeepSeek R1 Distill Llama 70B, and what is it best at?›
#98 of 133 in Chat & Knowledge.
Is DeepSeek R1 Distill Llama 70B better than Magistral Small 1?›
On GPQA Diamond · Epoch, DeepSeek R1 Distill Llama 70B scores 55.7 vs 56.1 — about the same.
How can I use DeepSeek R1 Distill Llama 70B?›
Through the API of DeepInfra.
What is the context window of DeepSeek R1 Distill Llama 70B?›
131K tokens.
When was DeepSeek R1 Distill Llama 70B released? Is there a newer version?›
Released 2025-01-20. It is the current version in its line.
Is DeepSeek R1 Distill Llama 70B open source?›
Yes, the weights are open (MIT licence).
Can I run DeepSeek R1 Distill Llama 70B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can DeepSeek R1 Distill Llama 70B read images?›
No, it accepts text only.
What is DeepSeek R1 Distill Llama 70B?›
DeepSeek R1 Distill Llama 70B is an open-weights language model from DeepSeek, released Jan 2025.
What are the best alternatives to DeepSeek R1 Distill Llama 70B?›
GPT-5.6 Luna (cheaper), Qwen3.6 27B (open weights).