Mercury 3 vs Qwen3.5 397B
Change modelsMercury 3Inception
Qwen3.5 397BAlibaba
IndexAll categories, out of 100.
53.1 ±7.9 best
53.0 ±2.4
GapIndex points below the top.
Top
−0.1 · no clear gap
CoverageIndex weight that has results.
75%
100% best
SpeedOutput tokens per second.
295/s best
53/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-10
2026-02
Input / 1MUS dollars per 1M input tokens.
$0.10 best
$0.60
Output / 1MUS dollars per 1M output tokens.
$0.35 best
$3.60
1K chats2K tokens in, 500 out, each.
$0.375 best
$3.00
1K documents20K tokens in, 1K out, each.
$2.35 best
$15.60
1K agent tasks200K tokens in, 20K out, each.
$27.00 best
$192.00
ContextMaximum tokens in one request.
1M best
128K
AgenticMulti-step tasks with tools.
47.9
52.4 best
CodingCode writing and repair.
63.5 best
52.5
ReasoningLogic problems and puzzles.
56.1 best
51.3
MultimodalTasks with images and text.
—
54.6
KnowledgeFacts and expert knowledge.
58.7 best
53.2
MultilingualTasks in many languages.
—
54.9
InstructionTasks with strict prompt rules.
59.8 best
54.8
MathMath problems.
—
52.6
Benchmarks
3 sharedOnly Mercury 3: 11Only Qwen3.5 397B: 67
Mercury 3
Qwen3.5 397B