Claude 3 Haiku vs Llama 3.1 70B Instruct
Wins 2 of 6 areas
Long documents · Following instructions
Wins 3 of 6 areas
Coding · Reasoning · Facts
Llama 3.1 70B Instruct wins more areas, narrowly.Claude 3 Haiku is better at long documents.
Scores updated · 30 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 70B Instruct2 of 2 tests
- FactsGetting facts right instead of making them up02Llama 3.1 70B Instruct2 of 2 tests
- ReasoningHard problems that need careful thinking01Llama 3.1 70B Instruct1 of 3 tests · 2 ties
- Long documentsFinding answers in very long texts10Claude 3 Haiku1 of 1 test
- Following instructionsDoing exactly what it is asked10Claude 3 Haiku1 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 1 tie
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.
Llama 3.1 70B Instruct costs 25% 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 Llama 3.1 70B Instruct pulls ahead
- Code for real scientific research problemsSciCode+8.1points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+3.5points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+2.3points ahead
Every test, side by side
All 30 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLlama 3.1 70B Instruct
- SciCodeLlama 3.1 70B Instruct by 8.118.626.7+8.1
- Terminal-Bench HardLlama 3.1 70B Instruct by 2.20.83+2.2
ReasoningLlama 3.1 70B Instruct
- GPQA DiamondLlama 3.1 70B Instruct by 3.537.440.9+3.5
- Humanity's Last Examtie4.14.5tie
- CritPttie00tie
FactsLlama 3.1 70B Instruct
- AA-Omniscience · Non-hallucinationLlama 3.1 70B Instruct by 2.319.521.8+2.3
- AA-Omniscience · AccuracyLlama 3.1 70B Instruct by 2.117.619.7+2.1
Following instructionsClaude 3 Haiku
- IFBenchClaude 3 Haiku by 1.736.134.4+1.7
Other results20 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.
- NarrativeQALlama 3.1 70B Instruct by 52.824.477.2+52.8
- HarmBenchClaude 3 Haiku by 44.491.346.9+44.4
- AIR-Bench 2024Claude 3 Haiku by 40.282.742.5+40.2
- naturalquestions_closedbookLlama 3.1 70B Instruct by 30.814.445.2+30.8
- MMLULlama 3.1 70B Instruct by 19.866.286+19.8
- MGSMLlama 3.1 70B Instruct by 11.875.186.9+11.8
- OpenBookQALlama 3.1 70B Instruct by 1083.893.8+10
- XSTestLlama 3.1 70B Instruct by 9.285.394.5+9.2
- BBHLlama 3.1 70B Instruct by 7.973.781.6+7.9
- simple_safety_testsClaude 3 Haiku by 7.510092.5+7.5
- anthropic_red_teamClaude 3 Haiku by 6.810093.2+6.8
- τ²-Bench Telecom (AA run)Claude 3 Haiku by 5.921.115.2+5.9
- ARC-ChallengeLlama 3.1 70B Instruct by 5.689.294.8+5.6
- AA-OmniscienceLlama 3.1 70B Instruct by 5.5-48.6-43.1+5.5
- GSM8KLlama 3.1 70B Instruct by 4.988.993.8+4.9
- HumanEvalLlama 3.1 70B Instruct by 4.675.980.5+4.6
- Artificial Analysis Coding IndexLlama 3.1 70B Instruct by 4.26.710.9+4.2
- AA Agentic IndexClaude 3 Haiku by 1.975.1+1.9
- DROPLlama 3.1 70B Instruct by 1.278.479.6+1.2
- AA IntelligenceLlama 3.1 70B Instruct by 15.66.6+1
Questions people ask
Which is better, Claude 3 Haiku or Llama 3.1 70B Instruct?
Llama 3.1 70B Instruct wins three of the six areas where both have results: coding, reasoning and facts. Claude 3 Haiku wins long documents and following instructions. They are level on agents.
Which is better for coding?
Llama 3.1 70B Instruct. It wins 2 of the 2 coding tests both models report; Claude 3 Haiku wins none.
Which is cheaper?
Claude 3 Haiku costs $0.25 per million input tokens and $1.25 per million output tokens; Llama 3.1 70B Instruct costs $0.56 and $0.56. That makes Llama 3.1 70B Instruct about 25% cheaper for the same work.
How do you compare the two?
We use the 30 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 20 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.