Muse Spark vs Beam 1.0 501B-A23B

Change models
Beam 1.0 501B-A23BReflection AI
IndexAll categories, out of 100.
59.0 ±3.5
59.3 ±4.6 best
GapIndex points below the top.
−0.3 · no clear gap
Top
CoverageIndex weight that has results.
95% best
85%
SpeedOutput tokens per second.
—
259/s
EloLMArena rating and rank.
1474
—
ReleasedMonth of release.
2026-04
2026-10
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
262K
1M best
AgenticMulti-step tasks with tools.
55.1
58.4 best
CodingCode writing and repair.
52.7
61.1 best
ReasoningLogic problems and puzzles.
58.7
62.2 best
MultimodalTasks with images and text.
62.0
—
KnowledgeFacts and expert knowledge.
66.8 best
56.4
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
59.0
62.3 best
MathMath problems.
59.5 best
54.8

Benchmarks

10 shared
Only Muse Spark: 37Only Beam 1.0 501B-A23B: 22
Muse Spark
Beam 1.0 501B-A23B
40.7% best45 / 216
34.7%64 / 216
78.0%70 / 211
78.2% best68 / 211
11.3%59 / 207
16.3% best46 / 207
49.6% best25 / 170
30.7%79 / 170
24.9%73 / 132
33.9% best54 / 132
77.4%36 / 114
81.6% best12 / 114
89.5%35 / 87
90.6% best27 / 87
52.4%64 / 83
66.1% best12 / 83
42.8% best25 / 55
36.0%31 / 55
74.8%18 / 25
80.5% best15 / 25