Claude Opus 4.7 vs GPT-5.4 mini
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents
Wins 0 of 8 areas
—
Claude Opus 4.7 is the stronger all-rounder.GPT-5.4 mini is cheaper.
Scores updated · 46 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. 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 software60Claude Opus 4.76 of 7 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own60Claude Opus 4.76 of 6 tests
- ReasoningHard problems that need careful thinking50Claude Opus 4.75 of 5 tests
- MathCompetition and research-level math30Claude Opus 4.73 of 3 tests
- FactsGetting facts right instead of making them up31Claude Opus 4.73 of 4 tests
- Images and chartsUnderstanding pictures, charts and video20Claude Opus 4.72 of 2 tests
- Long documentsFinding answers in very long texts10Claude Opus 4.71 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.
GPT-5.4 mini costs 83% 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 Opus 4.7 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+56.9points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+47.9points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+23.9points ahead
Where GPT-5.4 mini pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+14.7points ahead
Every test, side by side
All 46 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.7
- Terminal-Bench 2.1Claude Opus 4.7 by 23.983.159.2+23.9
- LMArena · WebDevClaude Opus 4.7 by 160 rating points15571397+160 rating
- LiveBench · CodingClaude Opus 4.7 by 10.582.171.6+10.5
- SWE-bench ProClaude Opus 4.7 by 9.964.354.4+9.9
- LiveBench · Agentic CodingClaude Opus 4.7 by 950.741.7+9
- SciCodeClaude Opus 4.7 by 2.454.552.1+2.4
- Terminal-Bench Hardtie51.552.3tie
AgentsClaude Opus 4.7
- MCP AtlasClaude Opus 4.7 by 19.677.357.7+19.6
- GDPValClaude Opus 4.7 by 1742.825.8+17
- AA IT-Bench SREClaude Opus 4.7 by 11.546.735.2+11.5
- τ-Bench V3 · BankingClaude Opus 4.7 by 934.625.6+9
- OSWorld-VerifiedClaude Opus 4.7 by 5.97872.1+5.9
- AA ApexAgentsClaude Opus 4.7 by 5.733.928.2+5.7
ReasoningClaude Opus 4.7
- ARC-AGI-2Claude Opus 4.7 by 56.975.818.9+56.9
- LiveBench · ReasoningClaude Opus 4.7 by 15.987.271.3+15.9
- Humanity's Last ExamClaude Opus 4.7 by 14.242.328.1+14.2
- GPQA DiamondClaude Opus 4.7 by 3.991.487.5+3.9
- CritPtClaude Opus 4.7 by 21210+2
FactsClaude Opus 4.7
- AA-Omniscience · Non-hallucinationClaude Opus 4.7 by 47.957.79.8+47.9
- SimpleQA VerifiedClaude Opus 4.7 by 22.351.729.4+22.3
- AA-Omniscience · AccuracyClaude Opus 4.7 by 11.448.937.5+11.4
- Vectara HHEM hallucination ratelower is betterGPT-5.4 mini by 6.5125.5+6.5
Images and chartsClaude Opus 4.7
- LMArena · VisionClaude Opus 4.7 by 68 rating points13131245+68 rating
- MMMU-ProClaude Opus 4.7 by 5.578.873.3+5.5
MathClaude Opus 4.7
- FrontierMath Tier 4Claude Opus 4.7 by 21.931.79.8+21.9
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.7 by 1970.251.2+19
- LiveBench · MathematicsClaude Opus 4.7 by 14.492.878.5+14.4
Following instructionsEven
- IFBenchGPT-5.4 mini by 14.758.673.3+14.7
- LiveBench · Instruction FollowingClaude Opus 4.7 by 6.966.759.8+6.9
Other results16 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 Opus 4.7 by 46.227.3-18.9+46.2
- ARC-AGI-1Claude Opus 4.7 by 28.39263.7+28.3
- frontiermath_tier_4_v1Claude Opus 4.7 by 20.822.92.1+20.8
- AA Agentic IndexClaude Opus 4.7 by 19.839.519.7+19.8
- Artificial Analysis Coding IndexClaude Opus 4.7 by 17.573.656.1+17.5
- AA IntelligenceClaude Opus 4.7 by 16.640.724.1+16.6
- ToolathlonClaude Opus 4.7 by 16.459.342.9+16.4
- vectara_avg_summary_lengthClaude Opus 4.7 by 94.4 rating points149.154.7+94.4 rating
- Terminal-Bench 2.0Claude Opus 4.7 by 9.469.460+9.4
- livebench_data_analysisClaude Opus 4.7 by 7.578.370.8+7.5
- Legal Agent BenchmarkClaude Opus 4.7 by 7.17.10+7.1
- livebench_languageClaude Opus 4.7 by 6.977.971+6.9
- vectara_factual_consistencyGPT-5.4 mini by 6.58894.5+6.5
- Finance Agent v2Claude Opus 4.7 by 6.151.545.4+6.1
- τ²-Bench Telecom (AA run)Claude Opus 4.7 by 5.388.683.3+5.3
- vectara_answer_rateGPT-5.4 mini by 298100+2
Questions people ask
Which is better, Claude Opus 4.7 or GPT-5.4 mini?
Claude Opus 4.7 wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and long documents. GPT-5.4 mini wins none, but costs 83% less. They are level on following instructions.
Which is better for coding?
Claude Opus 4.7. It wins 6 of the 7 coding tests both models report; GPT-5.4 mini wins none, and 1 is a tie.
Which is cheaper?
Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; GPT-5.4 mini costs $0.75 and $4.50. That makes GPT-5.4 mini about 83% cheaper for the same work.
How do you compare the two?
We use the 46 benchmark tests both models have published scores on. The verdict counts the 30 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.