DeepSeek-R1 0528-Qwen3-8B vs DeepSeek-R1-Distill-Llama-70B
Wins 2 of 5 areas
Reasoning · Long documents
Wins 3 of 5 areas
Coding · Facts · Following instructions
DeepSeek-R1-Distill-Llama-70B is the stronger all-rounder.DeepSeek-R1 0528-Qwen3-8B is better at reasoning.
Scores updated · 17 tests both models report · How we compare
Where each one wins
Tests won in each of the five 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.
- FactsGetting facts right instead of making them up02DeepSeek-R1-Distill-Llama-70B2 of 2 tests
- CodingWriting and fixing software01DeepSeek-R1-Distill-Llama-70B1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked01DeepSeek-R1-Distill-Llama-70B1 of 1 test
- ReasoningHard problems that need careful thinking10DeepSeek-R1 0528-Qwen3-8B1 of 3 tests · 2 ties
- Long documentsFinding answers in very long texts10DeepSeek-R1 0528-Qwen3-8B1 of 1 test
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.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where DeepSeek-R1-Distill-Llama-70B pulls ahead
- Code for real scientific research problemsSciCode+10.9points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+7.7points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+7points ahead
Where DeepSeek-R1 0528-Qwen3-8B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+21points ahead
- Reasons across sets of long documentsAA-LCR+5points ahead
Every test, side by side
All 17 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-R1-Distill-Llama-70B
- SciCodeDeepSeek-R1-Distill-Llama-70B by 10.920.431.3+10.9
- Terminal-Bench Hardtie1.51.5tie
ReasoningDeepSeek-R1 0528-Qwen3-8B
- GPQA DiamondDeepSeek-R1 0528-Qwen3-8B by 2161.240.2+21
- Humanity's Last Examtie5.95.1tie
- CritPttie00tie
FactsDeepSeek-R1-Distill-Llama-70B
- AA-Omniscience · AccuracyDeepSeek-R1-Distill-Llama-70B by 712.219.2+7
- AA-Omniscience · Non-hallucinationDeepSeek-R1-Distill-Llama-70B by 5.313.618.9+5.3
Long documentsDeepSeek-R1 0528-Qwen3-8B
- AA-LCRDeepSeek-R1 0528-Qwen3-8B by 51510+5
Following instructionsDeepSeek-R1-Distill-Llama-70B
- IFBenchDeepSeek-R1-Distill-Llama-70B by 7.719.927.6+7.7
Other results8 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.
- HMMT Feb. 2025DeepSeek-R1 0528-Qwen3-8B by 28.261.533.3+28.2
- τ²-Bench Telecom (AA run)DeepSeek-R1-Distill-Llama-70B by 21.9021.9+21.9
- AIME 2025DeepSeek-R1 0528-Qwen3-8B by 21.376.355+21.3
- AA-OmniscienceDeepSeek-R1-Distill-Llama-70B by 17.2-63.6-46.4+17.2
- AIME 2024DeepSeek-R1 0528-Qwen3-8B by 168670+16
- Artificial Analysis Coding IndexDeepSeek-R1-Distill-Llama-70B by 3.67.811.4+3.6
- LiveCodeBenchDeepSeek-R1 0528-Qwen3-8B by 360.557.5+3
- AA Intelligencetie8.17.9tie
Questions people ask
Which is better, DeepSeek-R1 0528-Qwen3-8B or DeepSeek-R1-Distill-Llama-70B?
DeepSeek-R1-Distill-Llama-70B wins three of the five areas where both have results: coding, facts and following instructions. DeepSeek-R1 0528-Qwen3-8B wins reasoning and long documents.
Which is better for coding?
DeepSeek-R1-Distill-Llama-70B. It wins 1 of the 2 coding tests both models report; DeepSeek-R1 0528-Qwen3-8B wins none, and 1 is a tie.
How do you compare the two?
We use the 17 benchmark tests both models have published scores on. The verdict counts the 9 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 8 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.