Claude Sonnet 4 vs Grok 3 Mini Reasoning
Wins 4 of 6 areas
Coding · Agents · Long documents · Following instructions
Wins 1 of 6 areas
Reasoning
Claude Sonnet 4 is the stronger all-rounder.Grok 3 Mini Reasoning is cheaper and better at reasoning.
Scores updated · 22 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 software10Claude Sonnet 41 of 2 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10Claude Sonnet 41 of 1 test
- Long documentsFinding answers in very long texts10Claude Sonnet 41 of 1 test
- Following instructionsDoing exactly what it is asked10Claude Sonnet 41 of 1 test
- ReasoningHard problems that need careful thinking01Grok 3 Mini Reasoning1 of 3 tests · 2 ties
- 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.
Grok 3 Mini Reasoning costs 96% 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 Claude Sonnet 4 pulls ahead
- Reasons across sets of long documentsAA-LCR+15points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+13.7points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+8.8points ahead
Where Grok 3 Mini Reasoning pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+3.3points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+1.4points ahead
Every test, side by side
All 22 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4
- Terminal-Bench HardClaude Sonnet 4 by 13.731.117.4+13.7
- SciCodetie4040.6tie
ReasoningGrok 3 Mini Reasoning
- GPQA DiamondGrok 3 Mini Reasoning by 1.477.779.1+1.4
- CritPttie0.30.6tie
- Humanity's Last Examtie10.711tie
FactsEven
- AA-Omniscience · AccuracyClaude Sonnet 4 by 7.622.715.1+7.6
- AA-Omniscience · Non-hallucinationGrok 3 Mini Reasoning by 3.370.974.2+3.3
Following instructionsClaude Sonnet 4
- IFBenchClaude Sonnet 4 by 8.854.745.9+8.8
Other results12 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 2025Grok 3 Mini Reasoning by 58.115.974+58.1
- τ²-Bench Telecom (AA run)Grok 3 Mini Reasoning by 25.864.690.4+25.8
- ARC-AGI-1Claude Sonnet 4 by 23.54016.5+23.5
- AIME 2025Grok 3 Mini Reasoning by 20.370.590.8+20.3
- livecodebench_hardGrok 3 Mini Reasoning by 18.532.651.1+18.5
- AA Agentic IndexGrok 3 Mini Reasoning by 13.617.631.2+13.6
- Artificial Analysis Coding IndexClaude Sonnet 4 by 12.437.625.2+12.4
- LiveCodeBenchGrok 3 Mini Reasoning by 11.968.580.4+11.9
- livecodebench_mediumGrok 3 Mini Reasoning by 9.97685.9+9.9
- AA-OmniscienceClaude Sonnet 4 by 70.2-6.8+7
- AA IntelligenceClaude Sonnet 4 by 4.318.914.6+4.3
- livecodebench_easytie98.898.1tie
Questions people ask
Which is better, Claude Sonnet 4 or Grok 3 Mini Reasoning?
Claude Sonnet 4 wins four of the six areas where both have results: coding, agents, long documents and following instructions. Grok 3 Mini Reasoning wins reasoning, and costs 96% less. They are level on facts.
Which is better for coding?
Claude Sonnet 4. It wins 1 of the 2 coding tests both models report; Grok 3 Mini Reasoning wins none, and 1 is a tie.
Which is cheaper?
Claude Sonnet 4 costs $3.00 per million input tokens and $15.00 per million output tokens; Grok 3 Mini Reasoning costs $0.30 and $0.50. That makes Grok 3 Mini Reasoning about 96% cheaper for the same work.
How do you compare the two?
We use the 22 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 12 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.