Loading benchmark catalogue
Compare performance benchmarks, coding intelligence, reasoning evaluations, context limits, and API economics between Claude Sonnet 5.5 (Anthropic) and Beam 1.0 501B A23B (Reflection AI).
| Metric / Capability | Claude Sonnet 5.5 | Beam 1.0 501B A23B |
|---|---|---|
| Developer / Organization | Anthropic | Reflection AI |
| Release Date | 2026-09-28 | 2026-10-16 |
| Deployment / Weights | Proprietary Cloud API | Open Weights |
| Context Window | 1M tokens (1,000,000 tokens) | 1.0M tokens (1,048,576 tokens) |
| Input Pricing (/1M Tokens) | $2 | Free (Open Weights) |
| Output Pricing (/1M Tokens) | $10 | Free (Open Weights) |
| Benchmark / Domain | Claude Sonnet 5.5 | Beam 1.0 501B A23B | Advantage |
|---|---|---|---|
Artificial Analysis Intelligence Index Composite index aggregating 10+ evaluations of general reasoning and knowledge. | 56.0% | — | — |
Artificial Analysis Coding Agent Index v2.0 Composite index of coding-agent systems (model + harness) across FrontierCode v1.1, SWE-Atlas-QnA, and Terminal-Bench v4. | 59.3% | — | — |
SWE-bench Pro Harder, contamination-resistant successor to SWE-bench spanning multiple languages. | 81.3% | 66.1% | Claude Sonnet 5.5 (+23%) |
Terminal-Bench 4.0 Sandboxed command-line tasks requiring shell and file-system operation. | 70.6% | — | — |
ARC-AGI-2 Novel visual grid puzzles requiring abstract rule induction. | 83.9% | — | — |
GPQA Diamond Graduate-level physics, chemistry, and biology questions written by domain experts. | — | 90.6% | — |
Humanity's Last Exam Expert-written questions across hundreds of disciplines designed to resist saturation. | 56.9% | 36.0% | Claude Sonnet 5.5 (+58%) |
FrontierMath v2 (Tiers 1-3) Unpublished, research-level mathematics problems across Tiers 1-3. | 88.8% | — | — |
Design Arena ELO Human-preference ELO across UI, 3D, data-visualization, and game-dev generations. | 1413 | 1227 | Claude Sonnet 5.5 (+15%) |
Claude Sonnet 5.5 leads across 3 evaluated benchmarks compared to 0 for Beam 1.0 501B A23B, showing notable advantages in SWE-bench Pro, Humanity's Last Exam, Design Arena ELO.
Beam 1.0 501B A23B has the larger context capacity of 1.0M tokens (1,048,576 tokens), compared to 1M tokens for Claude Sonnet 5.5.
Beam 1.0 501B A23B is more economical. Input tokens cost $Free (Open Weights)/M tokens compared to $2/M tokens for Claude Sonnet 5.5.
Claude Sonnet 5.5 is a proprietary cloud API. Beam 1.0 501B A23B is an open-weights model.
Analyze 130+ models side-by-side using interactive radar charts, cost-efficiency quadrant plots, and live dataset filters on the full NeuBench index.
Open Interactive Comparator