AI model reference · 2026
MiniMax M2.7
MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.
- Context window
- 204,800 tokens
- Maximum output
- 131,072 tokens
- Base provider API input
- $0.30 / 1M tokens
- Base provider API rate
- Published evidence
- 10 benchmarks, 2 Arena results
Overview
About MiniMax M2.7
A concise catalog overview. Technical limits and published evaluation evidence are listed separately below.
MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.
Published evaluations
Benchmark results
Only results mapped to this exact model and backed by a named source appear here. Different evaluation protocols are not treated as interchangeable.
| Benchmark | Score | Evaluation details | Source |
|---|---|---|---|
| MLE-Bench LiteA machine-learning engineering benchmark. MiniMax reports the average medal rate across three trials for M2.7. | 66.6% | Average medal rateVersion: MLE-Bench Lite · Attempts: 3 trialsProtocol note: Provider-reported average medal rate; detailed harness and reasoning settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| MM-ClawA multimodal agent benchmark reported by MiniMax, measuring accuracy under the provider's disclosed launch evaluation setup. | 62.7% | AccuracyVersion: MM-ClawProtocol note: Provider-reported launch result; detailed harness and reasoning settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| ToolathonA tool-use benchmark reported by MiniMax. The provider result is source-attributed but not treated as a controlled comparison without a matched protocol. | 46.3% | Success rateVersion: ToolathonProtocol note: Provider-reported launch result; detailed tool harness and reasoning settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| Multi-SWE-BenchA multi-language and multi-repository software-engineering benchmark reported by MiniMax, retained with its source-specific protocol identity. | 52.7% | ResolvedVersion: Multi-SWE-BenchProtocol note: Provider-reported launch result; detailed harness and task split are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| NL2RepoA repository-generation benchmark reported in the MiniMax M2.7 launch evaluation, with incomplete public harness disclosure in that source. | 39.8% | ScoreVersion: NL2RepoProtocol note: Provider-reported launch result; detailed harness and task split are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| SWE MultilingualA multilingual software-engineering benchmark reported by MiniMax. The launch source does not disclose a complete evaluation protocol. | 76.5% | ResolvedVersion: SWE MultilingualProtocol note: Provider-reported launch result; detailed harness and task split are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| SWE-Bench ProA contamination-resistant software-engineering benchmark with long-horizon tasks across multiple programming languages. Public and private splits are distinct protocols. | 56.2% | ResolvedVersion: SWE-Bench ProProtocol note: Provider-reported launch result; detailed split, harness, and reasoning settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| Terminal-Bench 2 (MiniMax report)The Terminal-Bench 2 result named in MiniMax's launch report. Its exact minor version and harness are not disclosed, so it must remain separate from versioned Terminal-Bench 2.0 and 2.1 rows. | 57.0% | Pass rateVersion: Terminal-Bench 2; exact minor version not disclosedProtocol note: Provider-reported launch result; kept separate from Terminal-Bench 2.0 and 2.1 because version and harness are not disclosedMethodology | MiniMaxProvider reported |
| VIBE-ProA provider-reported benchmark of long-horizon software-engineering work. The MiniMax launch source does not disclose enough harness detail for controlled cross-provider comparison. | 55.6% | ScoreVersion: VIBE-ProProtocol note: Provider-reported launch result; detailed harness and reasoning settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
| GDPval-AAAn Artificial Analysis GDPval evaluation reported as an Elo rating in MiniMax's official M2.7 launch post. | 1,495 | EloVersion: GDPval-AAProtocol note: Provider-reported Elo from the official launch table; detailed evaluation settings are not disclosed in the sourceMethodology | MiniMaxProvider reported |
Human preference
Arena results
Arena scores come from blind human preference votes. They are reported separately from task benchmarks and are not used as substitutes for missing benchmark results.
| Category | Score | Rank | Votes | Status |
|---|---|---|---|---|
| Arena TextBlind human preference score for text responses. | 1,416±4 | #115 | 55,513 | Published |
| Arena WebDevBlind human preference score for code and web development. | 1,398±6 | #63 | 12,473 | Published |
Technical reference
Specifications and API facts
Technical limits and reference provider prices come from the current catalog unless a reviewed external source is listed. Missing values stay marked as unavailable.
- Provider
- MiniMax
- Release date
- Mar 18, 2026
- Catalog added
- Mar 18, 2026
- Context window
- 204,800 tokens
- Maximum output
- 131,072 tokens
- Knowledge cutoff
- Not available in reviewed sources
- License
- Not available in reviewed sources
- Input types
- Text
- Output types
- Text
- Reasoning
- Supported
- Provider endpoint tool parameters
- Accepts tool parameters
- Base provider API input
- $0.30 / 1M tokens
- Base provider cached input
- $0.06 / 1M tokens
- Base provider API output
- $1.20 / 1M tokens
These are reference provider API rates, not Writingmate checkout charges. Access in Writingmate follows the allowances of your Writingmate plan.
Provenance
Sources and update status
Source links are attached to the facts and evaluations they support. Catalog-only values are not presented as independently verified claims.
- Arena TextArena AI · Retrieved Aug 9, 2026
- Arena WebDevArena AI · Retrieved Aug 9, 2026
- MiniMax M2.7: Early Echoes of Self-EvolutionMiniMax · Retrieved Aug 9, 2026
- OpenRouter model catalogOpenRouter
Evidence last updated Aug 9, 2026.
Catalog record updated Mar 18, 2026.
Catalog-added dates describe when a model entered the catalog, not necessarily its public release date.
Keep exploring
Related models and comparisons
Open another model’s detail page or compare it directly with MiniMax M2.7. Suggestions come from the current catalog.
Sakana Namazu
Sakana AI
Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts.
Solar Pro 4
Upstage
Solar Pro 4 is a large language model from Upstage.
Muse Glimmer 30B
Meta
Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware.
Muse Spark 1.2
Meta
Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks.
Qwen3.8 Max
Qwen
Qwen3.8 Max is the flagship model in Alibaba's Qwen3.8 series, the general-availability successor to the Qwen3.8 Max Preview.
DeepSeek V4 Flash 0731
DeepSeek
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
Try MiniMax M2.7 in Writingmate
Use this model, then compare its response with other AI models in the same workspace.