Beam 1.0
Beam 1.0 is a Reflection open-weights reasoning model, released 16 Oct 2026. Its best configuration ranks #246 of 775 on the BenchLeader Index at 54.6 ±8.6, in the upper half. The ± is the point: 423 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in long context (64) and lowest in agents & tools (43). At $1.05 per million tokens blended it is pricier than most ranked models. Output speed of 345 tokens per second puts it in the fastest quarter, with a first answer in 6.6 s. It has been measured at 5 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 16 Oct 2026.
- Blended price
- $1.05/M
- $0.700 in · $2.10 out
- Output speed
- 345 tok/s
- measured by Artificial Analysis
- First answer
- 6.58 s
- Context
- 1.0M
- Full answer
- 8.03 s
- median, reasoning included
- Cost per run
- $0.566
- one full Intelligence Index run
- Released
- 16 Oct 2026
- Overall index55
- Reasoning63
- Agents & tools43
- Knowledge62
- Long context64
- Composite50
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| low | – | – | 345 tok/s | 6.58 s | $0.0009 | Reasoning 36 |
| medium | – | – | 345 tok/s | 6.58 s | $0.0009 | Reasoning 39 |
| high | – | – | 345 tok/s | 6.58 s | $0.0009 | Reasoning 45 |
| xhigh | – | – | 345 tok/s | 6.58 s | $0.0009 | Reasoning 47 |
| not statedbest | 54.6 | #246 | 345 tok/s | 6.58 s | $0.0009 | Agents & tools 43 · Composite 50 · Knowledge 62 · Long context 64 · Reasoning 63 |
How its index has moved
17 Oct 2026 to 20 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| LiveBench 2.0 Reasoningnot in indexeffort not stated | – | – | – | – | 80.2%#38 | LiveBench | |
| Humanity's Last Exam (AA)not in indexeffort not stated | – | – | – | – | 34.7%#122 | Artificial Analysis | |
| ARC-AGI-1 | 45.0%#173 | 55.7%#160 | 67.7%#138 | 74.0%#127 | – | ARC Prize | |
| ARC-AGI-2 | 2.8%#184 | 6.4%#159 | 20.0%#133 | 23.5%#130 | – | ARC Prize | |
| CritPteffort not stated | – | – | – | – | 13.1%#102 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LiveBench 2.0 Codingnot in index | 83.9%#25 | LiveBench | |
| SciCode (AA)not in index | 48.5%#122 | Artificial Analysis | |
| Vibe Code Bench v1.1not in index | 66.4%#46 | Vals AI | |
| Terminal-Bench 4.0 (AA)not in index | 3.0%#106 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LiveBench Long Agenticnot in index | 17.3%#41 | LiveBench | |
| LiveBench Browsernot in index | 0.0%#45 | LiveBench | |
| GDPval-AA v2.1not in index | 33.9%#124 | Artificial Analysis | |
| Harvey's Legal Agent Benchmarknot in index | 4.6%#19 | Vals AI | |
| Legal Research Benchnot in index | 28.9%#50 | Vals AI | |
| Tax Agent Benchnot in index | 8.8%#57 | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | 4.0%#43 | Vals AI | |
| AutomationBenchnot in index | 48.5%#80 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 11.6%#114 | Artificial Analysis | |
| AA-Briefcase v1.1not in index | 976#51 | Artificial Analysis | |
| CWE-Bench (AA) | 49.2%#15 | Artificial Analysis | |
| DeepSec-Bench (AA) | 6.9%#17 | Artificial Analysis | |
| CyberGym end-to-end (AA) | 45.0%#8 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LiveBench 2.0 Data Analysisnot in index | 66.4%#37 | LiveBench | |
| LiveBench Researchnot in index | 72.8%#25 | LiveBench | |
| AA-Omniscience | 1.5#122 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 33.1%#163 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 52.8%#107 | Artificial Analysis |
Multimodal
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LiveBench Designnot in index | 67.9%#38 | LiveBench |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 78.0%#120 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LiveBench 2.0 | 55.5%#40 | LiveBench | |
| AA Intelligence Index v4.3.2 | 30.3#119 | Artificial Analysis | |
| Vals Indexnot in index | 43.4#20 | Vals AI | |
| AA Cyber Indexnot in index | 33.7#11 | Artificial Analysis |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0009 | – | 7.4 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0097 | – | 8.3 s |
| Code edit | 6,000 / 1,500 | $0.0074 | – | 10.9 s |
| Agentic coding session | 60,000 / 4,000 | $0.050 | – | 18.2 s |
| Structured extraction | 2,000 / 200 | $0.0018 | – | 7.2 s |
More from Reflection
See also
Data as of 21 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Beam 1.0: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/reflection-beam, data as of 21 Oct 2026.