Qwen3.8-27B vs Beam 1.0

Change models
Beam 1.0Unknown
IndexAll categories, out of 100.
58.3 ±2.1 best
58.1 ±8.6
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
100% best
65%
SpeedOutput tokens per second.
145/s
—
EloLMArena rating and rank.
1437
—
ReleasedMonth of release.
2026-08
—
Input / 1MUS dollars per 1M input tokens.
$0.06
—
Output / 1MUS dollars per 1M output tokens.
$1.96
—
1K chats2K tokens in, 500 out, each.
$1.10
—
1K documents20K tokens in, 1K out, each.
$3.16
—
1K agent tasks200K tokens in, 20K out, each.
$51.20
—
ContextMaximum tokens in one request.
1.02M
—
AgenticMulti-step tasks with tools.
55.9
60.0 best
CodingCode writing and repair.
59.6
62.5 best
ReasoningLogic problems and puzzles.
60.5 best
55.3
MultimodalTasks with images and text.
62.9
—
KnowledgeFacts and expert knowledge.
51.2
—
MultilingualTasks in many languages.
60.4
—
InstructionTasks with strict prompt rules.
67.2
—
MathMath problems.
55.8 best
53.1

Benchmarks

16 shared
Only Qwen3.8-27B: 85Only Beam 1.0: 5
Qwen3.8-27B
Beam 1.0
42.4% best57 / 137
23.5%63 / 137
87.5% best47 / 132
74.0%62 / 132
64.8%61 / 130
66.4% best60 / 130
38.8% best52 / 127
26.3%55 / 127
87.5% best53 / 116
79.1%56 / 116
36.1% best54 / 98
28.8%64 / 98
30.9% best34 / 89
8.8%76 / 89
76.6% best31 / 80
66.4%62 / 80
80.0%58 / 80
80.2% best56 / 80
16.0%47 / 55
18.0% best44 / 55
86.8% best19 / 52
83.9%26 / 52
75.9% best28 / 52
67.9%40 / 52
59.4%45 / 52
72.8% best27 / 52
27.9% best34 / 51
17.3%43 / 51
0.0%44 / 48
1.4% best40 / 48
IOICoding
39.1% best37 / 46
22.8%41 / 46