# GPT-5.6 Terra vs Mercury 2.5

> Side-by-side benchmark comparison for GPT-5.6 Terra and Mercury 2.5 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/gpt-5-6-terra-vs-mercury-2-5
- Last updated: September 8, 2026

- Shared sourced benchmarks: 1
- HTML indexing: indexable
- Ranking lane: BenchAlign v5

## Quick Verdict

Evidence is too thin for a benchmark verdict: GPT-5.6 Terra and Mercury 2.5 share only 1 sourced benchmark row.

## Summary

- Only 1 shared sourced benchmark row is available. That is not enough for a broad capability verdict, so this mirror does not name an overall winner.
- Mercury 2.5 has the lower published output-token price.
- GPT-5.6 Terra has the larger published context window at 1.05M.

## Model Snapshot

| Property | GPT-5.6 Terra | Mercury 2.5 |
|----------|----------|----------|
| Creator | OpenAI | Inception |
| Type | Proprietary | Proprietary |
| Reasoning | Reasoning | Reasoning |
| Context | 1.05M | 260K |
| Independent public score (not head-to-head) | 71.19 | null |
| Benchmarks Covered | 31 | 5 |
| Pricing (input/output) | $2.50 / $15.00 | $0.04 / $0.15 |

## Category Breakdown

### Agentic

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 60.2 (Supported · #20/152)
- Mercury 2.5 public-lane score: 44.5 (Estimated · #91/152)

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 3.0 | 20.8% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 87.4% | Coming soon | Coming soon |
| BrowseComp | 87.5% | Coming soon | Coming soon |
| OSWorld 2.0 | 50.2% | Coming soon | Coming soon |
| CyberGym | 81.8% | Coming soon | Coming soon |
| ExploitGym | 23.2% | Coming soon | Coming soon |
| Toolathlon | 53.1% | Coming soon | Coming soon |
| Terminal-Bench 2.1 (Vals) | 77.5% | Coming soon | Coming soon |
| τ³-bench results | Coming soon | 96.0% | Coming soon |
| DeepSearchQA | Coming soon | 34.0% | Coming soon |

### Coding

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 67.1 (Supported · #9/151)
- Mercury 2.5 public-lane score: 47.8 (Estimated · #73/151)

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Pro | 63.4% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 87.4% | Coming soon | Coming soon |
| deepSwe | 69.6% | Coming soon | Coming soon |
| FrontierCode 1.1 Extended | 55.8% | Coming soon | Coming soon |
| cursorBench32 | 64.9% | Coming soon | Coming soon |
| VulcanBench v3 | 87.0% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 85.9% | Coming soon | Coming soon |
| SWE-bench (Vals) | 95.4% | Coming soon | Coming soon |
| SciCode | Coming soon | 38% | Coming soon |

### Multimodal & Grounded

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 77.1 (#15/48)
- Mercury 2.5 public-lane score: Coming soon

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | 80.7% | Coming soon | Coming soon |
| MMMU-Pro w/ Python | 82% | Coming soon | Coming soon |

### Reasoning

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 64 (#11/18)
- Mercury 2.5 public-lane score: 67.7 (Unranked · 1 rankable row)

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-2 | 83.9% | Coming soon | Coming soon |
| ARC-AGI-3 | 0.8% | Coming soon | Coming soon |

### Knowledge

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 69.3 (Supported · #15/182)
- Mercury 2.5 public-lane score: 48.3 (Estimated · #94/182)

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 92.9% | Coming soon | Coming soon |
| GPQA-D | 92.9% | 79.0% | GPT-5.6 Terra |
| HLE-Verified | 51.1% | Coming soon | Coming soon |
| LABBench2 | 81.2% | Coming soon | Coming soon |
| HealthBench Professional | 57.7% | Coming soon | Coming soon |
| HealthBench Hard | 32.7% | Coming soon | Coming soon |
| GPQA Diamond (Vals) | 90.9% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 86.7% | Coming soon | Coming soon |

### Instruction Following

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 87.1 (#34/121)
- Mercury 2.5 public-lane score: 80.1 (#53/121)

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| IFBench | Coming soon | 77% | Coming soon |

### Mathematics

- Winner: Insufficient shared evidence
- GPT-5.6 Terra public-lane score: 97 (Unranked · 3 rankable rows)
- Mercury 2.5 public-lane score: Coming soon

| Benchmark | GPT-5.6 Terra | Mercury 2.5 | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath (legacy) | 84.9% | Coming soon | Coming soon |
| FrontierMath v2 (Tiers 1-3) | 84.900% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 68.300% | Coming soon | Coming soon |

## FAQ

### Is there enough evidence to pick between GPT-5.6 Terra and Mercury 2.5?

Not yet. The models share only 1 sourced benchmark row, so this mirror reports the available measurements and metadata without naming a broad winner.

### What can I compare safely on this page?

Published pricing, context windows, model type, and individual shared benchmark rows are directly comparable. Independent overall scores and non-overlapping benchmark coverage are labeled separately and should not be read as a controlled head-to-head result.

## Related Comparisons

- [GPT-5.6 Terra vs GPT-5.6 Sol](/compare/gpt-5-6-sol-vs-gpt-5-6-terra)
- [Mercury 2.5 vs GPT-5.6 Sol](/compare/gpt-5-6-sol-vs-mercury-2-5)
- [GPT-5.6 Terra vs GPT-5.6 Luna](/compare/gpt-5-6-luna-vs-gpt-5-6-terra)
- [Mercury 2.5 vs GPT-5.6 Luna](/compare/gpt-5-6-luna-vs-mercury-2-5)

## Explore More

- [GPT-5.6 Terra profile](/models/gpt-5-6-terra)
- [Mercury 2.5 profile](/models/mercury-2-5)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)