Beam 1.0 vs Mercury 3: Specs & Benchmark Comparison
Beam 1.0 is developed by Reflection AI, while Mercury 3 comes from Inception. Both were released in October 2026. Beam 1.0 has a published size of about 501 billion parameters; Inception has not disclosed the parameter count of Mercury 3.
The two models share 5 published benchmarks. Beam 1.0 leads on 5 of them. The widest gaps are on DeepSearchQA, where Beam 1.0 scores 80.5% against 71.3%; Tau3 Banking, where Beam 1.0 scores 36.1% against 30.2%. Averaged across everything we track, Beam 1.0 sits at 61.6% and Mercury 3 at 60.1%.
Mercury 3 is the cheaper API at $0.10 per million input tokens and $0.35 per million output tokens, roughly 4 times cheaper than Beam 1.0 at $0.40 and $1.20. Both accept a context window of about 1M tokens.
| Characteristic | Beam 1.0 | Mercury 3 |
|---|---|---|
| Company | Reflection AI | Inception |
| Release Date | October 16, 2026 | October 19, 2026 |
| Parameters | 501B | — |
| Multimodal | No | No |
| Context (input) | 1.0M | 1.0M |
| Context (output) | 131K | 131K |
| Input Price / 1M | $0.40 | $0.10 |
| Output Price / 1M | $1.20 | $0.35 |
| Average Score | 61.6% | 60.1% |
| Benchmarks | ||
| DeepSearchQA | 80.5% | 71.3% |
| Tau3 Banking | 36.1% | 30.2% |
| IFBench | 79.6% | 76.8% |
| GPQA | 90.6% | 89.2% |
| AA-LCR | 78.2% | 77.7% |
Visual Benchmark Comparison
Verdict
Mercury 3 leads in 1 out of 4 comparison categories.
Both models show comparable average scores: Beam 1.0 — 0.6, Mercury 3 — 0.6.
Mercury 3 is 3.6x cheaper: input $0.10/1M vs $0.40/1M tokens.
Same context size: 1M tokens.
Both models were released around the same time: 10/16/2026 and 10/19/2026.
More About These Models
Related Comparisons
Frequently Asked Questions
Which is better for coding — Beam 1.0 or Mercury 3?
Which model is cheaper — Beam 1.0 or Mercury 3?
Which has a larger context window — Beam 1.0 or Mercury 3?
The Beam 1.0 and Mercury 3 comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the Beam 1.0 or Mercury 3 page. See also the complete list of AI model comparisons.