Claude Sonnet 4.5 vs Llama 3.1 Instruct 405B
Wins 6 of 6 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 0 of 6 areas
—
Claude Sonnet 4.5 is the stronger all-rounder.
Scores updated · 29 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 software40Claude Sonnet 4.54 of 4 tests
- ReasoningHard problems that need careful thinking40Claude Sonnet 4.54 of 4 tests
- FactsGetting facts right instead of making them up20Claude Sonnet 4.52 of 2 tests
- Long documentsFinding answers in very long texts20Claude Sonnet 4.52 of 2 tests
- AgentsCarrying out multi-step tasks on its own10Claude Sonnet 4.51 of 1 test
- Following instructionsDoing exactly what it is asked10Claude Sonnet 4.51 of 1 test
Agents 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 Claude Sonnet 4.5 pulls ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+52.7points ahead
- Reasons across sets of long documentsAA-LCR+47points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+32.4points ahead
Where Llama 3.1 Instruct 405B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 29 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.5
- SWE-bench VerifiedClaude Sonnet 4.5 by 52.777.224.5+52.7
- SWE-bench ProClaude Sonnet 4.5 by 32.443.611.2+32.4
- Terminal-Bench HardClaude Sonnet 4.5 by 28.835.66.8+28.8
- SciCodeClaude Sonnet 4.5 by 15.845.729.9+15.8
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 31.983.451.5+31.9
- SimpleBenchClaude Sonnet 4.5 by 31.354.323+31.3
- Humanity's Last ExamClaude Sonnet 4.5 by 13.817.84+13.8
- CritPtClaude Sonnet 4.5 by 1.11.10+1.1
FactsClaude Sonnet 4.5
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 9.732.923.2+9.7
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.5 by 3.350.947.6+3.3
Long documentsClaude Sonnet 4.5
- AA-LCRClaude Sonnet 4.5 by 4772.325.3+47
- LongBench v2Claude Sonnet 4.5 by 25.761.836.1+25.7
Following instructionsClaude Sonnet 4.5
- IFBenchClaude Sonnet 4.5 by 18.357.339+18.3
Other results15 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.
- τ²-Bench Telecom (AA run)Claude Sonnet 4.5 by 59.178.119+59.1
- LiveCodeBenchClaude Sonnet 4.5 by 43.37127.7+43.3
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 37.652.114.5+37.6
- AIR-Bench 2024Claude Sonnet 4.5 by 31.289.858.6+31.2
- HarmBenchClaude Sonnet 4.5 by 29.39262.7+29.3
- AA-OmniscienceClaude Sonnet 4.5 by 17-0.1-17.1+17
- MMMLUClaude Sonnet 4.5 by 15.389.173.8+15.3
- MMLU-ProClaude Sonnet 4.5 by 14.287.573.3+14.2
- AA IntelligenceClaude Sonnet 4.5 by 13.420.77.3+13.4
- AA Agentic IndexClaude Sonnet 4.5 by 11.217.56.3+11.2
- MMLU-ReduxClaude Sonnet 4.5 by 9.495.686.2+9.4
- bbqClaude Sonnet 4.5 by 4.498.994.5+4.4
- anthropic_red_teamClaude Sonnet 4.5 by 1.798.296.5+1.7
- simple_safety_testsClaude Sonnet 4.5 by 1.210098.8+1.2
- XSTesttie96.495.9tie
Questions people ask
Which is better, Claude Sonnet 4.5 or Llama 3.1 Instruct 405B?
Claude Sonnet 4.5 wins all six areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Llama 3.1 Instruct 405B wins none.
Which is better for coding?
Claude Sonnet 4.5. It wins 4 of the 4 coding tests both models report; Llama 3.1 Instruct 405B wins none.
How do you compare the two?
We use the 29 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 15 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.