Ember-1 vs Mercury 3: Specs & Benchmark Comparison

Ember-1 is developed by Fireworks AI, while Mercury 3 comes from Inception. Ember-1 was released in September 2026, and Mercury 3 followed a month later in October 2026. Ember-1 has a published size of about 2.8 trillion parameters; Inception has not disclosed the parameter count of Mercury 3.

We have 4 benchmark results for Ember-1 and 12 benchmark results for Mercury 3, but they were measured on different benchmarks, so there is no like-for-like scoreboard.

Mercury 3 is the cheaper API at $0.10 per million input tokens and $0.35 per million output tokens, roughly 30 times cheaper than Ember-1 at $3 and $15. Both accept a context window of about 1M tokens.

CharacteristicEmber-1Mercury 3
CompanyFireworks AIInception
Release DateSeptember 23, 2026October 19, 2026
Parameters2.8T—
MultimodalNoNo
Context (input)1.0M1.0M
Context (output)—131K
Input Price / 1M$3.00$0.10
Output Price / 1M$15.00$0.35
Average Score78.8%60.1%

Verdict

Mercury 3 leads in 3 out of 4 comparison categories.

Overall Performance

Both models show comparable average scores: Ember-1 — 0.8, Mercury 3 — 0.6.

API Cost

Mercury 3 is 40.0x cheaper: input $0.10/1M vs $3.00/1M tokens.

Context Window

Mercury 3 supports a larger context: 1M vs 1M tokens.

Recency

Mercury 3 is newer: released 10/19/2026 vs 9/23/2026.

More About These Models

Related Comparisons

Frequently Asked Questions

Which is better for coding — Ember-1 or Mercury 3?
Direct comparison on the SWE-Bench benchmark is not available. We recommend reviewing other metrics on the comparison page.
Which model is cheaper — Ember-1 or Mercury 3?
Mercury 3 is cheaper for input: $0.10 per 1M tokens vs $3.00.
Which has a larger context window — Ember-1 or Mercury 3?
Mercury 3 supports a larger context: 1,048,576 tokens vs 1,040,000.

The Ember-1 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 Ember-1 or Mercury 3 page. See also the complete list of AI model comparisons.