The two are evenly matched.Claude Opus 4.6 is better at reasoning; Mercury 3 at following instructions.
Scores updated · 7 tests both models report · How we compare
Where each one wins
Tests won in each of the two 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.
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.
Mercury 3 costs 99% 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 Mercury 3 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+23.7points ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+23.2points ahead
Where Claude Opus 4.6 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+7points ahead
Every test, side by side
All 7 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningClaude Opus 4.6
- ARC-AGI-2Claude Opus 4.6 by 768.861.8+7
- ARC-AGI-3tie0.50.1tie
- GPQA Diamondtie89.689.2tie
Following instructionsMercury 3
- IFBenchMercury 3 by 23.753.176.8+23.7
- Multi-ChallengeMercury 3 by 23.237.160.4+23.2
Other results2 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.
- DeepSearchQA (F1)Claude Opus 4.6 by 2091.371.3+20
- ARC-AGI-1Claude Opus 4.6 by 4.59388.5+4.5
Questions people ask
Which is better, Claude Opus 4.6 or Mercury 3?
Claude Opus 4.6 and Mercury 3 each win one of the two areas where both have results. Claude Opus 4.6 wins reasoning; Mercury 3 wins following instructions. Mercury 3 costs 99% less.
Which is cheaper?
Claude Opus 4.6 costs $5.00 per million input tokens and $25.00 per million output tokens; Mercury 3 costs $0.10 and $0.35. That makes Mercury 3 about 99% cheaper for the same work.
How do you compare the two?
We use the 7 benchmark tests both models have published scores on. The verdict counts the 5 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 2 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.