Claude Sonnet 4.5 vs Granite 4.2 3B
Wins 4 of 6 areas
Coding · Agents · Reasoning · Long documents
Wins 1 of 6 areas
Following instructions
Claude Sonnet 4.5 is the stronger all-rounder.Granite 4.2 3B is cheaper and better at following instructions.
Scores updated · 20 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.
- ReasoningHard problems that need careful thinking30Claude Sonnet 4.53 of 3 tests
- AgentsCarrying out multi-step tasks on its own20Claude Sonnet 4.52 of 3 tests · 1 tie
- CodingWriting and fixing software21Claude Sonnet 4.52 of 3 tests
- Long documentsFinding answers in very long texts10Claude Sonnet 4.51 of 1 test
- Following instructionsDoing exactly what it is asked01Granite 4.2 3B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
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.
Granite 4.2 3B costs 99% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Sonnet 4.5 pulls ahead
- Reasons across sets of long documentsAA-LCR+48points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+41.9points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+27.5points ahead
Where Granite 4.2 3B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+22.8points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+17points ahead
- Recent programming contest problemsLiveCodeBench v6+5.7points ahead
Every test, side by side
All 20 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.5
- Terminal-Bench 2.1Claude Sonnet 4.5 by 41.955.813.9+41.9
- SciCodeClaude Sonnet 4.5 by 20.445.725.3+20.4
- LiveCodeBench v6Granite 4.2 3B by 5.76469.7+5.7
AgentsClaude Sonnet 4.5
- GDPValClaude Sonnet 4.5 by 20.520.50+20.5
- τ-Bench V3 · BankingClaude Sonnet 4.5 by 18.924.55.6+18.9
- Terminal-Bench 4.0tie00tie
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 27.583.455.9+27.5
- Humanity's Last ExamClaude Sonnet 4.5 by 11.217.86.6+11.2
- CritPtClaude Sonnet 4.5 by 1.11.10+1.1
FactsEven
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 23.732.99.2+23.7
- AA-Omniscience · Non-hallucinationGranite 4.2 3B by 22.850.973.7+22.8
Following instructionsGranite 4.2 3B
- IFBenchGranite 4.2 3B by 1757.374.3+17
Other results7 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.
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 34.652.117.5+34.6
- MMLU-ProClaude Sonnet 4.5 by 20.488.267.8+20.4
- AA-OmniscienceClaude Sonnet 4.5 by 14.6-0.1-14.7+14.6
- AA IntelligenceClaude Sonnet 4.5 by 11.620.79.1+11.6
- AIME25 no toolsClaude Sonnet 4.5 by 9.78878.3+9.7
- AIME 2025Claude Sonnet 4.5 by 8.78778.3+8.7
- HMMT 2025Claude Sonnet 4.5 by 7.974.666.7+7.9
Questions people ask
Which is better, Claude Sonnet 4.5 or Granite 4.2 3B?
Claude Sonnet 4.5 wins four of the six areas where both have results: coding, agents, reasoning and long documents. Granite 4.2 3B wins following instructions, and costs 99% less. They are level on facts.
Which is better for coding?
Claude Sonnet 4.5. It wins 2 of the 3 coding tests both models report; Granite 4.2 3B wins 1.
Which is cheaper?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens; Granite 4.2 3B costs $0.03 and $0.12. That makes Granite 4.2 3B about 99% cheaper for the same work.
How do you compare the two?
We use the 20 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 7 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.