vs
0:9
overall capability wins (0 ties across 9 scored). MiniMax M2.5 leads overall.
- Reasoning0:6
- Coding0:2
- Pricing3:0
- Specs0:1
Verdict
MiniMax M2.5 leads coding (widest gap: +217.5 on WebDev Elo); MiniMax M2.5 offers ~7.8x larger context (1M vs 128K); MiniMax M2.5 ships open weights (Modified MIT).
- MiniMax M2.5 leads coding (widest gap: +217.5 on WebDev Elo)
- MiniMax M2.5 offers ~7.8x larger context (1M vs 128K)
- MiniMax M2.5 ships open weights (Modified MIT)
- Mercury 2 is ~1.4x cheaper on a blended token basis than MiniMax M2.5
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Mercury 2
- $0.44
- MiniMax M2.5
- $0.60
Mercury 2 is about 1.4x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Inception Labs
MiniMax
Family
Other
Other
License
Proprietary
Modified MIT
Open weights
No
Yes
Release
Sep 1, 2026
Feb 15, 2026
Knowledge cutoff
-
-
API / provider
Inception Labs
MiniMax
Modalities
text → text
text → text
Specs
Context window
128K
1M
Max output
-
64K
Parameters
—
—
Pricing
Input $/1M
$0.25
$0.30
Output $/1M
$0.75
$1.20
Blended $/1M (3∶1)
$0.38
$0.52
Speed
tok/s
-
85
TTFT (s)
-
0.35
Reasoning
MMLU-Pro
—
—
GPQA Diamond
77.0
84.8
Humanity's Last Exam
17.1
20.5
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
21.2
26.2
AA-LCR v1.1
43.7
73.3
CritPt
0.8
1.1
MMMU-Pro
—
—
IFBench
69.8
71.6
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
75.8
SWE-bench Pro
—
—
SWE-bench Multilingual
—
68.3
LiveCodeBench
—
—
Terminal-Bench 2.1
27.3
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
37.7
42.6
CursorBench
—
—
SWE-Rebench
—
47.7
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,167
1,384
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
1,390
| Metric | Mercury 2 | MiniMax M2.5 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Inception Labs | MiniMax | — |
| Family | Other | Other | — |
| License | Proprietary | Modified MIT | — |
| Open weights | No | Yes | — |
| Release | Sep 1, 2026 | Feb 15, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Inception Labs | MiniMax | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 128K | 1M | B +872K |
| Max output | - | 64K | — |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $0.25 | $0.30 | A +$0.05 |
| Output $/1M | $0.75 | $1.20 | A +$0.45 |
| Blended $/1M (3∶1) | $0.38 | $0.52 | A +$0.15 |
| Speed | |||
| tok/s | - | 85 | — |
| TTFT (s) | - | 0.35 | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 77.0 | 84.8 | B +7.8 pts |
| Humanity's Last Exam | 17.1 | 20.5 | B +3.4 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 21.2 | 26.2 | B +5.0 pts |
| AA-LCR v1.1 | 43.7 | 73.3 | B +29.6 pts |
| CritPt | 0.8 | 1.1 | B +0.3 pts |
| MMMU-Pro | — | — | — |
| IFBench | 69.8 | 71.6 | B +1.8 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 75.8 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | 68.3 | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 27.3 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 37.7 | 42.6 | B +4.9 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | 47.7 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,167 | 1,384 | B +218 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | 1,390 | — |
Mercury 2
- Context
- 128K
- Parameters
- —
- Price
- $0.25 / $0.75
- Speed
- —
- Modalities
- text → text
- License
- Proprietary
Inception Labs' diffusion-based (non-autoregressive) reasoning LLM, API-only with a 128K context window and high-throughput inference on Blackwell GPUs.
MiniMax M2.5
- Context
- 1M / 64K out
- Parameters
- —
- Price
- $0.30 / $1.20
- Speed
- 85 tok/s · 0.35s TTFT
- Modalities
- text → text
- License
- Modified MIT
Open-ish MiniMax MoE with long context and competitive Chinese-market pricing.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.