MiniMax M2
MiniMaxhuggingface.co/MiniMaxAI/MiniMax-M2
MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows. With 10 billion activated parameters (230 billion total), it delivers near-frontier intelligence across general reasoning,...
Updated 20 min ago · first seen 11 Sept 2026
model_01M294WW9PDCZ47SBF9A0Q2TF9
- Parameters
- 228.7B
- T2 · 6 h ago
- Context
- 204.8K tokens
- T2 · 6 h ago
- Released
- 22 Oct 2025
- T2 · 1 h ago
- License
- other
- T2 · 5 h ago
Specification
- Release date
- 22 Oct 2025
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
- Status
- deprecated
Source:Artificial AnalysisT2observed 4 h agomedium
- Openness
- open-weights
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
- License
- other
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Architecture
- MiniMaxM2ForCausalLM
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Parameters
- 228.7B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
- Context window
- 204.8K tokens
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Max output
- 131.1K tokens
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Modalities
- text
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Input modalities
- text
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Output modalities
- text
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Quantization format
- fp8
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- File size
- 230.1 GB
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Hugging Face repo
- MiniMaxAI/MiniMax-M2
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Pipeline tag
- text-generation
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
- Downloads
- 233,394
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 20 min agomedium
- Likes
- 1,503
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 20 min agomedium
- Gated
- No
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Hf inference providers
- featherless-ai, novita
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Is quantized
- Yes
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Last modified
- 2025-12-23T08:37:43+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
- Library name
- transformers
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- License name
- modified-mit
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- License url
- https://github.com/MiniMax-AI/MiniMax-M2/blob/main/LICENSE
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Aa median output tokens per second
- 80.7
Source:Artificial AnalysisT2observed 4 h agomedium
- Downloads all time
- 2,685,351
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
- Model type
- minimax_m2
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Openrouter id
- minimax/minimax-m2
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Supported parameters
- frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Tags
- transformers, safetensors, minimax_m2, text-generation, custom_code, fp8
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
- Weights dtype
- BF16, F32, F8_E4M3
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 h agomedium
Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →
Provenance
Attributed facts
36
Source tiers
T236
Freshest observation
20 min ago
Conflicts
None
Modalities
- Modalities
- text
- Input
- text
- Output
- text
Capabilities
Tool calling
Yes
OpenRouter public model & pricing listing · T2
Structured output
Yes
OpenRouter public model & pricing listing · T2
Reasoning
Yes
OpenRouter public model & pricing listing · T2
Vision
Unavailable
Audio
Unavailable
Fine-tuning available
Unavailable
- Context window
- 204.8K tokens
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
- Max output
- 131.1K tokens
Source:OpenRouter public model & pricing listingT2observed 6 h agomedium
| Benchmark | Score | Metric | Config | Evaluated | Source |
|---|---|---|---|---|---|
| τ²-bench | 86.84% | pass^1 | aa_slug=minimax-m2 · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| Artificial Analysis Intelligence Index | 18.63 | index | aa_slug=minimax-m2 · version=4.3 · estimated=true | — | artificialanalysis.aiT2 |
| GPQA | 77.68% | accuracy | aa_slug=minimax-m2 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| Humanity's Last Exam | 13.67% | accuracy | aa_slug=minimax-m2 · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| IFBench | 72.31% | accuracy | aa_slug=minimax-m2 · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| SWE-bench Verified | 61% | resolved | date=2025-11-24 · board=Verified · system=mini-SWE-agent · model_tag=minimax-m2 | 24 Nov 2025 | swebench.comT2 |
| Terminal-Bench | 25.76% | accuracy | aa_slug=minimax-m2 · variant=hard · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
Scores are reported as published, with their evaluation configuration (harness, prompting, judge). Results with different configs are not directly comparable — see methodology.
Current prices
| Provider | Input / 1M | Output / 1M | Cached in | Batch in / out | Context | Observed | Source |
|---|---|---|---|---|---|---|---|
| MiniMax APIminimax/minimax-m2 | $0.255 | $1.02 | — | — | 204.8K | 2 h ago | openrouter.aiT2 |
USD per 1M tokens as published by each provider (USD). Rows are append-only: every change is kept in the history below.
Available through 1
Price history
Input price · USD / 1M tokens
Output price · USD / 1M tokens
- MiniMax API
Memory need = bytes per parameter (4-bit ≈ 0.5 × 1.15 overhead, 8-bit 1.0, fp16 2.0) + a KV-cache allowance. Not a measurement.
| Hardware | Quantization | Memory | Est. need | Fits |
|---|---|---|---|---|
| NVIDIA DGX B200 | 4bit | 1,440 GB | 132.0 GB | Yes |
| Apple M3 Ultra | 4bit | — | 132.0 GB | Yes |
| Mac Studio (Apple M5 Ultra) | 4bit | — | 132.0 GB | Yes |
| AMD Instinct MI325X | 4bit | 256 GB | 132.0 GB | Yes |
| AMD Instinct MI300X | 4bit | 192 GB | 132.0 GB | Yes |
| Apple M2 Ultra | 4bit | — | 132.0 GB | Yes |
| NVIDIA B200 | 4bit | 180 GB | 132.0 GB | Yes |
| NVIDIA H200 | 4bit | 141 GB | 132.0 GB | Yes |
| NVIDIA H200 NVL | 4bit | 141 GB | 132.0 GB | Yes |
| Apple M1 Ultra | 4bit | — | 132.0 GB | No |
| Apple M3 Max | 4bit | — | 132.0 GB | No |
| Apple M4 Max | 4bit | — | 132.0 GB | No |
| Mac Studio (Apple M5 Max) | 4bit | — | 132.0 GB | No |
| MacBook Pro (Apple M5 Max) | 4bit | — | 132.0 GB | No |
| NVIDIA DGX Spark | 4bit | 128 GB | 132.0 GB | No |
| Apple M2 Max | 4bit | — | 132.0 GB | No |
| NVIDIA H100 NVL | 4bit | 94 GB | 132.0 GB | No |
| NVIDIA A100 80GB | 4bit | 80 GB | 132.0 GB | No |
| NVIDIA H100 SXM | 4bit | 80 GB | 132.0 GB | No |
| Apple M1 Max | 4bit | — | 132.0 GB | No |
| Apple M4 Pro | 4bit | — | 132.0 GB | No |
| Mac mini (Apple M5 Pro) | 4bit | — | 132.0 GB | No |
| MacBook Pro (Apple M5 Pro) | 4bit | — | 132.0 GB | No |
| Apple M3 Pro | 4bit | — | 132.0 GB | No |
| Apple M1 Pro | 4bit | — | 132.0 GB | No |
| Apple M2 Pro | 4bit | — | 132.0 GB | No |
| Apple M4 | 4bit | — | 132.0 GB | No |
| iMac (Apple M4) | 4bit | — | 132.0 GB | No |
| Mac mini (Apple M6) | 4bit | — | 132.0 GB | No |
| MacBook Air (Apple M5) | 4bit | — | 132.0 GB | No |
| MacBook Pro (Apple M5) | 4bit | — | 132.0 GB | No |
| NVIDIA GeForce RTX 5090 | 4bit | 32 GB | 132.0 GB | No |
| Apple M2 | 4bit | — | 132.0 GB | No |
| Apple M3 | 4bit | — | 132.0 GB | No |
| NVIDIA GeForce RTX 3090 | 4bit | 24 GB | 132.0 GB | No |
| NVIDIA GeForce RTX 4090 | 4bit | 24 GB | 132.0 GB | No |
| Apple M1 | 4bit | — | 132.0 GB | No |
Ancestors 0
None recorded.
This model
MiniMax M2
228.7B params
Quantizations 3
Descendants 0
None recorded.
Papers 3
- arXiv:2504.07164Active35
- arXiv:2509.06501Active35
- arXiv:2509.13160Active35
Repositories 0
No repositories linked yet.
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history · Supported parameters
Supported parameterssupported_parameters1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p | → current | current | OpenRouter public model & pricing listingT2 | medium | deterministic |
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
MiniMax M2 scores 61% on SWE-bench Verified
swebench_leaderboardMiniMax M2 scores 13.67% on Humanity's Last Exam
artificial_analysisMiniMax M2 scores 18.63 on Artificial Analysis Intelligence Index
artificial_analysisMiniMax M2: release date changed from 2025-10-22 to 2025-10-23
Release date22 Oct 2025→23 Oct 2025openrouterMiniMax M2: release date changed from 2025-10-23 to 2025-10-22
Release date23 Oct 2025→22 Oct 2025huggingfaceMiniMax API lists MiniMax M2 at $0.255 in / $1.02 out per 1M tokens
openrouter
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/models?author=MiniMaxAI&p=0&sort=downloads | listing | T2· Quality secondary | 20 min ago | 6 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/MiniMaxAI/MiniMax-M2/raw/main/README.md | model_card | T2· Quality secondary | 34 min ago | 1 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/MiniMaxAI/MiniMax-M2 | model_page | T2· Quality secondary | 1 h ago | 4 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/cerebras/MiniMax-M2-REAP-172B-A10B | model_page | T2· Quality secondary | 1 h ago | 4 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/cerebras/MiniMax-M2-REAP-162B-A10B | model_page | T2· Quality secondary | 1 h ago | 4 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/cerebras/MiniMax-M2-REAP-139B-A10B | model_page | T2· Quality secondary | 1 h ago | 4 |
| OpenRouter public model & pricing listing | openrouter.ai/api/v1/models | catalogue | T2· Quality secondary | 2 h ago | 4 |
| SWE-bench leaderboards | swebench.com/ | leaderboard | T2· Quality secondary | 4 h ago | 1 |
| Artificial Analysis | artificialanalysis.ai/leaderboards/models | leaderboard | T2· Quality secondary | 4 h ago | 1 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.