MDL
DeepSeek R1 Distill Llama 70B
DeepSeek
Verified against Artificial Analysis · Jul 13, 2026
Capability
49
percentile index
Price /M
$0.7 / $1.05
in / out
Context
128K
tokens
Avg score
35.7
11 benchmarks
AA Intelligence
9.9
index
Speed
28 tok/s
TTFT 0.5s
Released
Jan 20, 2025
About
DeepSeek R1 Distill Llama 70B is a open-weights model in the DeepSeek family from DeepSeek. Benchmarked on 11 evals, averaging 35.7.
Benchmark scores · 11
avg 35.7 — higher is betterThe research behind this model
all papers mentioning it →ThoughtFold: Folding Reasoning Chains via Introspective Preference Learning▲ 25 on HF · Jun 2, 2026N-GRPO: Embedding-Level Neighbor Mixing for Enhanced Policy Optimization▲ 24 on HF · Jun 9, 2026ESPO: Early-Stopping Proximal Policy Optimization▲ 19 on HF · May 28, 2026Information-Aware KV Cache Compression for Long Reasoning▲ 11 on HF · Jun 25, 2026RoPE-Aware Bit Allocation for KV-Cache Quantization▲ 8 on HF · Jun 23, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed28 tok/s
Latency (TTFT)0.5s
First answer1m 12s
Price in / out$0.7 / $1.05 per 1M
Blended price$0.787 / 1M
Math index53.7
Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →
History
full ledger →Capability over time
Append-only ledger — every observed change to this model's numbers.
API providers
Artificial Analysis$1.05/M