Claude Sonnet 4 vs o4-mini
Wins 3 of 7 areas
Coding · Facts · Long documents
Wins 3 of 7 areas
Agents · Reasoning · Images and charts
The two are evenly matched.Claude Sonnet 4 is better at coding and facts; o4-mini at images and charts and agents.
Scores updated · 42 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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.
- Images and chartsUnderstanding pictures, charts and video02o4-mini2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02o4-mini2 of 2 tests
- ReasoningHard problems that need careful thinking01o4-mini1 of 4 tests · 3 ties
- CodingWriting and fixing software21Claude Sonnet 42 of 3 tests
- FactsGetting facts right instead of making them up21Claude Sonnet 42 of 3 tests
- Long documentsFinding answers in very long texts10Claude Sonnet 41 of 1 test
- Following instructionsDoing exactly what it is asked11Even1 each
Long documents rests on a single test.
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.
o4-mini costs 69% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where o4-mini pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+39.3points ahead
- Real work tasks from 44 professionsGDPVal+15.7points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+14points ahead
Where Claude Sonnet 4 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+51.4points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+15.9points ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+14.1points ahead
Every test, side by side
All 42 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 15.931.115.2+15.9
- SciCodeo4-mini by 6.54046.5+6.5
- SWE-bench VerifiedClaude Sonnet 4 by 4.672.768.1+4.6
Agentso4-mini
- BrowseCompo4-mini by 39.312.251.5+39.3
- GDPValo4-mini by 15.79.725.4+15.7
Reasoningo4-mini
- Humanity's Last Examo4-mini by 5.810.716.5+5.8
- GPQA Diamondtie77.778.4tie
- CritPttie0.30.6tie
- ARC-AGI-2tie5.96.1tie
FactsClaude Sonnet 4
- AA-Omniscience · Non-hallucinationClaude Sonnet 4 by 51.470.919.5+51.4
- Vectara HHEM hallucination ratelower is betterClaude Sonnet 4 by 8.310.318.6+8.3
- AA-Omniscience · Accuracyo4-mini by 2.122.724.8+2.1
Images and chartso4-mini
- MMMU-Proo4-mini by 7.461.869.2+7.4
- MMMUo4-mini by 7.274.481.6+7.2
- LMArena · Visiontie11911194tie
Following instructionsEven
- Multi-ChallengeClaude Sonnet 4 by 14.157.143+14.1
- IFBencho4-mini by 1454.768.7+14
Other results24 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.
- AA-OmniscienceClaude Sonnet 4 by 35.90.2-35.7+35.9
- AIME 2025o4-mini by 22.270.592.7+22.2
- ARC-AGI-1o4-mini by 18.74058.7+18.7
- AA Agentic Indexo4-mini by 18.517.636.1+18.5
- livecodebench_hardo4-mini by 16.332.648.9+16.3
- Artificial Analysis Coding IndexClaude Sonnet 4 by 1237.625.6+12
- TAU-bench (airline)Claude Sonnet 4 by 10.86049.2+10.8
- livecodebench_mediumo4-mini by 10.27686.2+10.2
- AIR-Bench 2024Claude Sonnet 4 by 9.888.378.5+9.8
- τ²-Bench Telecom (AA run)Claude Sonnet 4 by 964.655.6+9
- LiveCodeBencho4-mini by 8.968.577.4+8.9
- TAU-bench (retail)Claude Sonnet 4 by 8.780.571.8+8.7
- vectara_factual_consistencyClaude Sonnet 4 by 8.389.781.4+8.3
- frontiermath_tier_4_v1o4-mini by 6.306.3+6.3
- bbqClaude Sonnet 4 by 2.696.694+2.6
- AA IntelligenceClaude Sonnet 4 by 2.218.916.7+2.2
- vectara_avg_summary_lengthClaude Sonnet 4 by 18.1 rating points145.8127.7+18.1 rating
- Aider-PolyglotClaude Sonnet 4 by 1.870.768.9+1.8
- anthropic_red_teamClaude Sonnet 4 by 199.298.2+1
- HarmBenchClaude Sonnet 4 by 19897+1
- XSTesttie96.697.4tie
- vectara_answer_ratetie98.699.2tie
- livecodebench_easytie98.898.8tie
- simple_safety_teststie100100tie
Questions people ask
Which is better, Claude Sonnet 4 or o4-mini?
Claude Sonnet 4 and o4-mini each win three of the seven areas where both have results. Claude Sonnet 4 wins coding, facts and long documents; o4-mini wins agents, reasoning and images and charts. o4-mini costs 69% less. They are level on following instructions.
Which is better for coding?
Claude Sonnet 4. It wins 2 of the 3 coding tests both models report; o4-mini wins 1.
Which is cheaper?
Claude Sonnet 4 costs $3.00 per million input tokens and $15.00 per million output tokens; o4-mini costs $1.10 and $4.40. That makes o4-mini about 69% cheaper for the same work.
How do you compare the two?
We use the 42 benchmark tests both models have published scores on. The verdict counts the 18 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 24 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.