Beam 1.0
Beam 1.0 is an open-weights language model from Reflection, released October 5, 2026. It has 501B parameters (mixture-of-experts, 23B active). The context window is 1M tokens. On the Epoch Capabilities Index it scores 150.9, rank 66 of 256 rated models. API price: $0 input / $0 output per million tokens.
Check it against your own PC →
Can you run Beam 1.0 locally?
Even with 32 GB of VRAM and 32 GB of RAM it is too large to run at a usable speed on a single consumer GPU. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Too large | – | – | – | – |
| 8 GB VRAM | Too large | – | – | – | – |
| 12 GB VRAM | Too large | – | – | – | – |
| 16 GB VRAM | Too large | – | – | – | – |
| 24 GB VRAM | Too large | – | – | – | – |
| 32 GB VRAM | Too large | – | – | – | – |
Ready-made quantizations on Hugging Face: bartowski/Beam-1.0-501B-A23B-GGUF · unsloth/Beam-1.0-501B-A23B-GGUF · Baekpica/Beam-1.0-501B-A23B-VQ-GGUF · tablinum/Lite-Beam-1.0-501B-A23B-IQS_XXS · nbKPS/beam-reap-255b-gguf · DevQuasar/reflection.Beam-1.0-501B-A23B-GGUF
Benchmark results
| Benchmark (Epoch AI) | Score |
|---|---|
| ProofBench | 18.0% |
Links
Hugging Face: reflection/Beam-1.0-501B-A23B · Epoch AI benchmarks
Models with a similar score
Muse Spark 1.1 · Mistral Large 4 · Qwen3.7-Max · GPT-5.2 · DeepSeek V4 Flash 0731 · GLM-5.3-Flash · Muse Spark · GLM-5.2