Reflection AIopen-weight
Released: 2026-10-16Artificial AnalysisBeam 1.0
Beam 1.0 from Reflection, tracked on the Artificial Analysis Intelligence Index.
Intelligence Index
17.7/ 100
Artificial Analysis Intelligence Index (independent composite)Throughput Speed
274tok/sec
Output streaming throughputTime to First Answer
6.7s
Time to first answer token, including any thinkingCost to Run the Index
0.28USD
Artificial Analysis cost of the full Index runMeasured results · Artificial Analysis
Independent evaluations of Beam 1.0. Values are reported as published; nothing here is estimated or derived. Snapshot 2026-11-15.
- Intelligence Index
- 17.7
- Cost to run the Index
- $0.28
- Time per task
- 2.1 min
- Omniscience index
- 2.3
- Reasoning model
- Yes
- Cached-input price
- $0.04 / 1M
Reasoning & knowledge
- AA-Omniscience Accuracy30.7%
- CritPt13.1%
- CritPt v2 (Hard)26.5%
- HLE-Diamond12.8%
- Humanity's Last Exam34.7%
Coding & terminal agents
- FrontierCode 1.124.3%
- FrontierCode-AA 2.05.5%
- SciCode48.5%
- Terminal-Bench 4.03.0%
- Terminal-Bench-Science 0.10.5%
Agentic work
- AA-Briefcase v1.1 (Elo)976 Elo
- AutomationBench-AA48.5%
- GDPval-AA (normalized Elo)1177 Elo
Long context & documents
Multimodal & science
- Code2Video-AA (Elo)675 Elo
Health & safety
Headline Evaluation Breakdown
Six independently measured evaluations (Artificial Analysis)Reasoning · CritPt v2 (Hard)26.5%
Coding & agents · Terminal-Bench 4.03.0%
Long context · AA-LCR78.0%
Knowledge · Humanity's Last Exam34.7%
Sourced Benchmark Evaluations (15)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| FrontierCode | Active | 24.3% | 98 | 2026-11-15 | independent | Source ↗ |
| Humanity's Last Exam | Active | 34.7% | 93 | 2026-11-15 | independent | Source ↗ |
| SciCode | Active | 48.5% | 93 | 2026-11-15 | independent | Source ↗ |
| Terminal-Bench 4.0 | Active | 3% | 93 | 2026-11-15 | independent | Source ↗ |
| Terminal-Bench-Science 0.1 | Active | 0.5% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-LCR (Long Context Reasoning) | Active | 78% | 93 | 2026-11-15 | independent | Source ↗ |
| CritPt | Active | 13.1% | 93 | 2026-11-15 | independent | Source ↗ |
| CritPt v2 (Hard) | Active | 26.5% | 93 | 2026-11-15 | independent | Source ↗ |
| HLE-Diamond | Active | 12.8% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-Omniscience (Accuracy) | Active | 30.7% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-Omniscience (Non-Hallucination) | Active | 59% | 93 | 2026-11-15 | independent | Source ↗ |
| FrontierCode-AA 2.0 | Active | 5.5% | 93 | 2026-11-15 | independent | Source ↗ |
| GDP.pdf | Active | 11.6% | 93 | 2026-11-15 | independent | Source ↗ |
| MLCR-AA (Medical Long Context Reasoning) | Active | 17.8% | 93 | 2026-11-15 | independent | Source ↗ |
| AutomationBench-AA | Active | 48.5% | 93 | 2026-11-15 | independent | Source ↗ |
Local Hardware Execution (Ollama)Hardware Compatible
# 1. Pull and execute model locally with Ollama
ollama run reflection-beam-1-0
# 2. Or invoke via local OpenAI-compatible endpoint
curl http://localhost:11434/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "reflection-beam-1-0",
"messages": [{"role": "user", "content": "Analyze reasoning chains on AIME 2026"}]
}'Evaluation & Sourcing Notes
Scores listed for Beam 1.0 represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.