Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
Mercury 2 is clearly ahead on the aggregate, 65 to 60. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Mercury 2's sharpest advantage is in agentic, where it averages 63.7 against 52.3. The single biggest benchmark swing on the page is OSWorld-Verified, 62 to 49. Kimi K2.5 does hit back in instruction following, so the answer changes if that is the part of the workload you care about most.
Kimi K2.5 is also the more expensive model on tokens at $0.50 input / $2.80 output per 1M tokens, versus $0.25 input / $0.75 output per 1M tokens for Mercury 2. That is roughly 3.7x on output cost alone. Mercury 2 is the reasoning model in the pair, while Kimi K2.5 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use.
Pick Mercury 2 if you want the stronger benchmark profile. Kimi K2.5 only becomes the better choice if instruction following is the priority or you would rather avoid the extra latency and token burn of a reasoning model.
Mercury 2
63.7
Kimi K2.5
52.3
Mercury 2
41.1
Kimi K2.5
38.9
Mercury 2
68.3
Kimi K2.5
64.6
Mercury 2
80.1
Kimi K2.5
71.7
Mercury 2
57.2
Kimi K2.5
57.2
Mercury 2
84
Kimi K2.5
85
Mercury 2
79.7
Kimi K2.5
79.8
Mercury 2
80.9
Kimi K2.5
78.7
Mercury 2 is ahead overall, 65 to 60. The biggest single separator in this matchup is OSWorld-Verified, where the scores are 62 and 49.
Mercury 2 and Kimi K2.5 are effectively tied for knowledge tasks here, both landing at 57.2 on average.
Mercury 2 has the edge for coding in this comparison, averaging 41.1 versus 38.9. Inside this category, HumanEval is the benchmark that creates the most daylight between them.
Mercury 2 has the edge for math in this comparison, averaging 80.9 versus 78.7. Inside this category, AIME 2023 is the benchmark that creates the most daylight between them.
Mercury 2 has the edge for reasoning in this comparison, averaging 80.1 versus 71.7. Inside this category, MuSR is the benchmark that creates the most daylight between them.
Mercury 2 has the edge for agentic tasks in this comparison, averaging 63.7 versus 52.3. Inside this category, OSWorld-Verified is the benchmark that creates the most daylight between them.
Mercury 2 has the edge for multimodal and grounded tasks in this comparison, averaging 68.3 versus 64.6. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.
Kimi K2.5 has the edge for instruction following in this comparison, averaging 85 versus 84. Inside this category, IFEval is the benchmark that creates the most daylight between them.
Kimi K2.5 has the edge for multilingual tasks in this comparison, averaging 79.8 versus 79.7. Inside this category, MGSM is the benchmark that creates the most daylight between them.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.