Reflection AI model product
Beam 1.0
4B active parameters per token, and a 1M-token context window.
Updated Oct 23, 2026. Default version: Beam 1.0
LLMBoard Score73.3Beam 1.0
Coverage20%22 benchmark families
Context window1MTokens
Official input priceN/AOfficial price unavailable
Beam 1.0 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Beam 1.0 Benchmark Results
Benchmark scores for Beam 1.0.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkLongBench v2 | Score67.50% | Rank01 | Participants19 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkAIME 2026 | Score97.70% | Rank02 | Participants27 | Percentile96.15% | EvidenceC | Evaluated |
| BenchmarkAA-LCR | Score78.20% | Rank03 | Participants27 | Percentile92.31% | EvidenceC | Evaluated |
| BenchmarkAA-Omniscience Index | Score10.80 points | Rank03 | Participants5 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkCritPT | Score16.30% | Rank03 | Participants7 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkTau3 Banking | Score36.10% | Rank03 | Participants16 | Percentile86.67% | EvidenceC | Evaluated |
| BenchmarkLiveCodeBench | Score88.30% | Rank04 | Participants76 | Percentile96.00% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam (with tools) | Score51.10% | Rank05 | Participants5 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkSWE Atlas - Codebase QnA | Score31.80% | Rank07 | Participants7 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Pro | Score66.10% | Rank07 | Participants63 | Percentile90.32% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score81.60% | Rank07 | Participants120 | Percentile94.96% | EvidenceC | Evaluated |
| BenchmarkMCP Atlas | Score80.80% | Rank08 | Participants39 | Percentile81.58% | EvidenceC | Evaluated |
| BenchmarkDeepSearchQA | Score80.50% | Rank09 | Participants13 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkIFBench | Score79.60% | Rank09 | Participants46 | Percentile82.22% | EvidenceC | Evaluated |
| BenchmarkSciCode | Score49.20% | Rank09 | Participants31 | Percentile73.33% | EvidenceC | Evaluated |
| BenchmarkSWE-bench Multilingual | Score78.30% | Rank12 | Participants50 | Percentile77.55% | EvidenceC | Evaluated |
| BenchmarkAutomationBench | Score39.30% | Rank13 | Participants27 | Percentile53.85% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Bash Reliability | Score4.30% | Rank16 | Participants44 | Percentile65.12% | EvidenceA | Evaluated |
| BenchmarkTerminal-Bench 2.1 | Score80.10% | Rank23 | Participants48 | Percentile53.19% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score90.60% | Rank24 | Participants256 | Percentile90.98% | EvidenceC | Evaluated |
| BenchmarkBrowseComp | Score76.80% | Rank28 | Participants70 | Percentile60.87% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Leaderboard | Score-4.35% | Rank30 | Participants44 | Percentile32.56% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Praise Complaint | Score-5.78% | Rank30 | Participants44 | Percentile32.56% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Steerability | Score-5.35% | Rank32 | Participants44 | Percentile27.91% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Task Outcome Explicit | Score-9.43% | Rank34 | Participants44 | Percentile23.26% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Tool Hallucination | Score-1.18% | Rank36 | Participants44 | Percentile18.60% | EvidenceA | Evaluated |
| BenchmarkDeepSWE 1.1 | Score43.30% | Rank45 | Participants51 | Percentile12.00% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam | Score36.00% | Rank47 | Participants109 | Percentile57.41% | EvidenceC | Evaluated |
| BenchmarkLM Arena Webdev | Score1,436.86 rating | Rank57 | Participants107 | Percentile47.17% | EvidenceA | Evaluated |
| BenchmarkLM Arena Text Factuality | Score1,441.00 rating | Rank70 | Participants131 | Percentile46.92% | EvidenceA | Evaluated |
Beam 1.0 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenaagent bash reliability | Categoryoverall | Rank16 | Rating / score0.04 | VotesN/A | Observations87.1K | Result date |
| Arenaagent | Categoryoverall | Rank30 | Rating / score-0.04 | VotesN/A | Observations654K | Result date |
| Arenaagent praise complaint | Categoryoverall | Rank30 | Rating / score-0.06 | VotesN/A | Observations2.2K | Result date |
| Arenaagent steerability | Categoryoverall | Rank32 | Rating / score-0.05 | VotesN/A | Observations41K | Result date |
| Arenatext factuality | Categorycoding | Rank32 | Rating / score1,519.07 | Votes1,012 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry mathematical | Rank33 | Rating / score1,486.92 | Votes251 | ObservationsN/A | Result date |
| Arenaagent task outcome explicit | Categoryoverall | Rank34 | Rating / score-0.09 | VotesN/A | Observations4.8K | Result date |
| Arenaagent tool hallucination | Categoryoverall | Rank36 | Rating / score-0.01 | VotesN/A | Observations519K | Result date |
| Arenatext | Categoryindustry mathematical | Rank38 | Rating / score1,475.28 | Votes251 | ObservationsN/A | Result date |
| Arenatext style control | Categorymath | Rank40 | Rating / score1,472.26 | Votes233 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry software and it services | Rank41 | Rating / score1,493.14 | Votes1,460 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry business and management and financial operations | Rank44 | Rating / score1,458.89 | Votes702 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryinstruction following | Rank44 | Rating / score1,453.10 | Votes1,542 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryhard prompts english | Rank45 | Rating / score1,487.59 | Votes1,037 | ObservationsN/A | Result date |
| Arenatext | Categorycoding | Rank48 | Rating / score1,472.54 | Votes1,019 | ObservationsN/A | Result date |
| Arenatext | Categoryfrench | Rank48 | Rating / score1,461.46 | Votes257 | ObservationsN/A | Result date |
| Arenatext style control | Categorycoding | Rank48 | Rating / score1,504.11 | Votes1,019 | ObservationsN/A | Result date |
| Arenatext | Categorymath | Rank49 | Rating / score1,458.84 | Votes233 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-react | Rank53 | Rating / score1,442.51 | Votes1,850 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryenglish | Rank54 | Rating / score1,457.75 | Votes1,588 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryhard prompts | Rank55 | Rating / score1,469.77 | Votes2,676 | ObservationsN/A | Result date |
| Arenatext | Categoryinstruction following | Rank56 | Rating / score1,430.34 | Votes1,557 | ObservationsN/A | Result date |
| Arenawebdev | Categoryoverall | Rank57 | Rating / score1,436.86 | Votes2,930 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev | Rank57 | Rating / score1,436.86 | Votes2,930 | ObservationsN/A | Result date |
| Arenatext style control | Categoryinstruction following | Rank58 | Rating / score1,439.00 | Votes1,557 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-html | Rank59 | Rating / score1,436.18 | Votes405 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry entertainment and sports and media | Rank60 | Rating / score1,414.92 | Votes1,172 | ObservationsN/A | Result date |
| Arenatext style control | Categoryfrench | Rank60 | Rating / score1,469.73 | Votes257 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry software and it services | Rank61 | Rating / score1,460.81 | Votes1,460 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry software and it services | Rank62 | Rating / score1,484.90 | Votes1,460 | ObservationsN/A | Result date |
Beam 1.0 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $0.25 input, $1 output per 1M via Kilo Gateway
- Tracked offerings
- 6
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderReflection | Provider model IDBeam-1.0-501B-A23B | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context1M | Updated |
| ProviderVercel AI Gateway | Provider model IDreflection/beam | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context1M | Updated |
| ProviderKilo Gateway | Provider model IDreflection/beam-1.0-501b-a23b | Regionglobal | Input / 1M$0.25 | Output / 1M$1 | Context1M | Updated |
| ProviderOpenRouter | Provider model IDreflection/beam-1.0-501b-a23b | Regionglobal | Input / 1M$0.25 | Output / 1M$1 | Context1M | Updated |
| ProviderCoreWeave | Provider model IDreflection/Beam-1.0-501B-A23B | Regionglobal | Input / 1M$0.30 | Output / 1M$1 | Context1M | Updated |
| ProviderTogether AI | Provider model IDreflection/Beam-1.0-501B-A23B | Regionglobal | Input / 1M$0.40 | Output / 1M$1.2 | Context1M | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Beam 1.0 Runtime Performance
Provider-specific output speed and catalog latency for Beam 1.0. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Beam 1.0 Specifications
Technical details for the model's default version.
- Version
- Beam 1.0
- Released
- Oct 16, 2026
- Knowledge cutoff
- Jun 30, 2026
- Parameters
- 501.3B
- Context window
- 1M
- Max output
- 131.1K
- Inputs
- text
- Outputs
- text
- Open weights
- Yes
- License
- Apache 2.0
Beam 1.0 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionBeam 1.0 | Released | LLMBoard73.25 | Parameters501.3B | Context1M | Max output131.1K | Open weightsYes | LicenseApache 2.0 |
Models similar to Beam 1.0
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MA
Kimi K2.6
Moonshot AI
AN
Claude Mythos 5.1
Anthropic
OP
GPT-5.4
OpenAI
AC
Qwen3.8 27B
Alibaba Cloud / Qwen Team
MA
Mistral Large 4
Mistral AI
XA
Grok 4.6
xAI
What is Beam 1.0?
Key information about Beam 1.0 and its available data.
0 is Reflection AI’s first open-weight model, designed for coding, long-horizon agentic work, and tool use. 0 license.
Data as of 2026-10-22.
FAQ
Common questions about Beam 1.0.
When was Beam 1.0 released?
Beam 1.0's default version was released on Oct 16, 2026.
How much does Beam 1.0 cost?
No official standard PAYG price is currently available for Beam 1.0. The lowest tracked third-party offer starts at $0.25 input and $1 output via Kilo Gateway.
Who created Beam 1.0?
Beam 1.0 was created by Reflection AI.
What is the context window for Beam 1.0?
The default version has a 1M token context window.
Is Beam 1.0 open weight?
Yes. The default version is marked as open weight under Apache 2.0.
How many API providers offer Beam 1.0?
6 provider offerings are linked to the default version.