Llama 3.1 8B Instruct vs Llama 3.1 Nemotron Instruct 70B
Wins 1 of 6 areas
Long documents
Wins 2 of 6 areas
Coding · Following instructions
Llama 3.1 Nemotron Instruct 70B wins more areas, narrowly.Llama 3.1 8B Instruct is better at long documents.
Scores updated · 18 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 software02Llama 3.1 Nemotron Instruct 70B2 of 2 tests
- Following instructionsDoing exactly what it is asked01Llama 3.1 Nemotron Instruct 70B1 of 1 test
- Long documentsFinding answers in very long texts10Llama 3.1 8B Instruct1 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 1 tie
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
- FactsGetting facts right instead of making them up11Even1 each
Agents, 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 three tests each model wins by the widest margin. Scores are out of 100.
Where Llama 3.1 Nemotron Instruct 70B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+20.6points ahead
- Code for real scientific research problemsSciCode+10.1points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+9.3points ahead
Where Llama 3.1 8B Instruct pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+28.3points ahead
- Reasons across sets of long documentsAA-LCR+9.7points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1.1points ahead
Every test, side by side
All 18 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLlama 3.1 Nemotron Instruct 70B
- SciCodeLlama 3.1 Nemotron Instruct 70B by 10.113.223.3+10.1
- Terminal-Bench HardLlama 3.1 Nemotron Instruct 70B by 3.70.84.5+3.7
ReasoningEven
- GPQA DiamondLlama 3.1 Nemotron Instruct 70B by 20.625.946.5+20.6
- Humanity's Last ExamLlama 3.1 8B Instruct by 1.15.34.2+1.1
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationLlama 3.1 8B Instruct by 28.35728.7+28.3
- AA-Omniscience · AccuracyLlama 3.1 Nemotron Instruct 70B by 9.38.517.8+9.3
Long documentsLlama 3.1 8B Instruct
- AA-LCRLlama 3.1 8B Instruct by 9.7188.3+9.7
Following instructionsLlama 3.1 Nemotron Instruct 70B
- IFBenchLlama 3.1 Nemotron Instruct 70B by 2.128.630.7+2.1
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.
- Arena HardLlama 3.1 Nemotron Instruct 70B by 48.636.485+48.6
- ARC-ChallengeLlama 3.1 8B Instruct by 14.283.469.2+14.2
- GSM8KLlama 3.1 Nemotron Instruct 70B by 11.679.891.4+11.6
- AA-OmniscienceLlama 3.1 8B Instruct by 9.9-30.9-40.8+9.9
- τ²-Bench Telecom (AA run)Llama 3.1 Nemotron Instruct 70B by 6.716.423.1+6.7
- TruthfulQALlama 3.1 Nemotron Instruct 70B by 5.852.858.6+5.8
- Artificial Analysis Coding IndexLlama 3.1 Nemotron Instruct 70B by 5.45.410.8+5.4
- AA Intelligencetie6.96.9tie
Questions people ask
Which is better, Llama 3.1 8B Instruct or Llama 3.1 Nemotron Instruct 70B?
Llama 3.1 Nemotron Instruct 70B wins two of the six areas where both have results: coding and following instructions. Llama 3.1 8B Instruct wins long documents. They are level on agents, reasoning and facts.
Which is better for coding?
Llama 3.1 Nemotron Instruct 70B. It wins 2 of the 2 coding tests both models report; Llama 3.1 8B Instruct wins none.
How do you compare the two?
We use the 18 benchmark tests both models have published scores on. The verdict counts the 10 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.