Skip to content
aiqbench

Reflection · model passport

Reflection: Beam 1.0 501B A23B

beam-1-0-501b-a23b
  • Open weights
  • Apache 2.0 licence
  • 501.3B params
  • 1,048,576 context
  • Released 2026-10-07(HF repo created)
2/27axes covered
Best percentile
50.4WebDev Arena
Cohort standing
45.9median 60.7

Thin coverage — read with caution

This model appears on only 2 of 27 benchmark axes considered. The overall standing has been shrunk toward the neutral 50 to reflect the small sample, and no weaknesses have been claimed — a low percentile on 2 axes is a coverage artifact, not a measured weakness.

What the data says

Benchmarked on 2 of 27 tracked axes. Every figure below is the model's own published result on that board.

On only 2 board(s), a ranked list of strengths would be false authority — the per-axis figures are shown below instead, without a verdict.

Uneven: 21.3 percentile points between its best and worst axis — 50.4 pct on WebDev Arena versus 29.1 pct on Agent Arena.

  • WebDev Arena50.4 pct1,438Verify →
  • Agent Arena29.1 pct-0.037Verify →
  • OverallNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked
  • CodingNot benchmarked
  • Instruction followingNot benchmarked
  • Hard promptsNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

Repositoryreflection/Beam-1.0-501B-A23B
Parameters501.3B
ArchitectureBeamForCausalLM
LicenceApache 2.0

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "reflection/Beam-1.0-501B-A23B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDreflection/beam-1.0-501b-a23b
Context length1,048,576 tokens
Max output943,718 tokens
Input price$1 per 1M tokens
Cached input price$0.10 per 1M tokens
Output price$3 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin-pPresence penaltyReasoningReasoning effortRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-kTop logprobsTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"reflection/beam-1.0-501b-a23b","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • beam-1.0-501b-a23b stands at 45.9 (raw mean 39.7) — measured on only 2 of 27 axes, so the standing is shrunk toward the neutral 50.

    standingtemplateVerify →

  • beam-1.0-501b-a23b stands at 45.9 versus a cohort median of 60.7 across 171 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →