Inference
EN
Reflection AI

Beam 1.0 501B-A23B FP8

Beam 1.0 501B-A23B FP8 is a reasoning model from Reflection AI in the Beam family. 17 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-10
RankedEnough results for an index.
IndexAll categories, out of 100.59.2±5.4Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.7 counted55.4101 / 218
CodingCode writing and repair.3 counted61.463 / 209
ReasoningLogic problems and puzzles.3 counted64.348 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.2 counted60.363 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted64.4—
MathMath problems.1 counted55.8—

Results

17 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
AIME 2026Math98.0%2 / 34
LongBench v2Reasoning67.6%2 / 25
SWE-bench ProCoding65.9%13 / 83
Software Engineering Benchmark VerifiedCoding80.2%19 / 114
Instruction Following BenchmarkInstruction80.7%10 / 57
Critical Physics TasksReasoning17.4%41 / 214
Artificial Analysis Long Context ReasoningReasoning79.4%51 / 217
MCP AtlasAgentic81.5%13 / 52
Scientific Code BenchmarkCoding48.3%11 / 35
GPQA Diamond · BenchLMKnowledge90.3%28 / 84
Terminal-Bench 2.1 · BenchLMAgentic78.9%39 / 86
Humanity's Last Exam without toolsKnowledge36.0%30 / 53
BrowseCompAgentic76.7%36 / 63
DeepSearchQAAgentic79.6%15 / 24
τ³-Bench Tool-Agent-User EvaluationAgentic36.8%20 / 32
Humanity's Last Exam with toolsAgentic51.1%25 / 36
DeepSWEAgentic43.0%55 / 59
SWE-Atlas Codebase QnACoding32.2%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index60.3 → 59.2
23 Oct 202612 Nov 2026

Questions

How good is Beam 1.0 501B-A23B FP8?
Its index is 59.2 out of 100. That is place 77 of 283 ranked models.
What is Beam 1.0 501B-A23B FP8 best at?
Reasoning: 64.3 out of 100, place 48 of 227.
What is the context window of Beam 1.0 501B-A23B FP8?
1M tokens.
Is Beam 1.0 501B-A23B FP8 open source?
Yes. Its weights are published.
When was Beam 1.0 501B-A23B FP8 released?
October 2026.

Badge

Beam 1.0 501B-A23B FP8 on Inference360: place 77 of 283, index 59.2
HTML
<a href="https://inference360.ai/models/reflection-beam-fp8"><img src="https://inference360.ai/badges/reflection-beam-fp8.svg" alt="Beam 1.0 501B-A23B FP8 on Inference360: place 77 of 283, index 59.2" height="20"></a>
Markdown
[![Beam 1.0 501B-A23B FP8 on Inference360: place 77 of 283, index 59.2](https://inference360.ai/badges/reflection-beam-fp8.svg)](https://inference360.ai/models/reflection-beam-fp8)

Compare with

Qwen3.8-27B59.2GPT-5.259.0Muse Spark59.8Beam 1.0 501B-A23B NVFP458.5Beam 1.0 501B-A23B60.2

More from Reflection AI

All models
Beam 1.0 501B-A23B60.2Beam 1.0 501B-A23B NVFP458.5