DeepSeek-R1 vs Llama 3.3 70B Instruct
Wins 4 of 6 areas
Coding · Agents · Reasoning · Long documents
Wins 1 of 6 areas
Following instructions
DeepSeek-R1 is the stronger all-rounder.Llama 3.3 70B Instruct is cheaper and better at following instructions.
Scores updated · 26 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software30DeepSeek-R13 of 3 tests
- ReasoningHard problems that need careful thinking20DeepSeek-R12 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10DeepSeek-R11 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10DeepSeek-R11 of 1 test
- Following instructionsDoing exactly what it is asked01Llama 3.3 70B Instruct1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
Long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Llama 3.3 70B Instruct costs 76% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where DeepSeek-R1 pulls ahead
- Reasons across sets of long documentsAA-LCR+42points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+21points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+14.2points ahead
Where Llama 3.3 70B Instruct pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+8.1points ahead
Every test, side by side
All 26 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-R1
- Terminal-Bench 2.1DeepSeek-R1 by 14.219.14.9+14.2
- SciCodeDeepSeek-R1 by 12.338.326+12.3
- Terminal-Bench HardDeepSeek-R1 by 3.16.13+3.1
ReasoningDeepSeek-R1
- GPQA DiamondDeepSeek-R1 by 2170.849.8+21
- Humanity's Last ExamDeepSeek-R1 by 4.98.53.6+4.9
- CritPttie0.60tie
FactsEven
- AA-Omniscience · AccuracyDeepSeek-R1 by 11.530.518.9+11.5
- Vectara HHEM hallucination ratelower is betterLlama 3.3 70B Instruct by 7.211.34.1+7.2
- AA-Omniscience · Non-hallucinationtie9.79.8tie
Following instructionsLlama 3.3 70B Instruct
- IFBenchLlama 3.3 70B Instruct by 8.13947.1+8.1
Other results13 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- LiveCodeBenchDeepSeek-R1 by 30.263.533.3+30.2
- vectara_avg_summary_lengthDeepSeek-R1 by 28.993.564.6+28.9
- AA-OmniscienceDeepSeek-R1 by 22-32.2-54.2+22
- MATHDeepSeek-R1 by 20.397.377+20.3
- τ²-Bench Telecom (AA run)Llama 3.3 70B Instruct by 15.211.426.6+15.2
- MMLU-ProDeepSeek-R1 by 15.18468.9+15.1
- Artificial Analysis Coding IndexDeepSeek-R1 by 12.724.611.9+12.7
- SimpleQADeepSeek-R1 by 9.230.120.9+9.2
- IFEvalLlama 3.3 70B Instruct by 8.883.392.1+8.8
- vectara_factual_consistencyLlama 3.3 70B Instruct by 7.288.795.9+7.2
- MMLUDeepSeek-R1 by 4.890.886+4.8
- AA IntelligenceDeepSeek-R1 by 3.711.47.7+3.7
- vectara_answer_rateLlama 3.3 70B Instruct by 2.59799.5+2.5
Questions people ask
Which is better, DeepSeek-R1 or Llama 3.3 70B Instruct?
DeepSeek-R1 wins four of the six areas where both have results: coding, agents, reasoning and long documents. Llama 3.3 70B Instruct wins following instructions, and costs 76% less. They are level on facts.
Which is better for coding?
DeepSeek-R1. It wins 3 of the 3 coding tests both models report; Llama 3.3 70B Instruct wins none.
Which is cheaper?
DeepSeek-R1 costs $2.00 per million input tokens and $4.00 per million output tokens; Llama 3.3 70B Instruct costs $0.71 and $0.72. That makes Llama 3.3 70B Instruct about 76% cheaper for the same work.
How do you compare the two?
We use the 26 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.