MiniMax M3
MiniMax
- 上下文視窗
- 1M 1,000,000 token
- 輸入 / 輸出
- 影像, 文字, 影片 → 文字
時間軸
- · 權重版本 · 參考連結
輸入/輸出價格 · USD / 百萬 token
| 模型 | 參考連結 | 輸入/輸出價格 |
|---|---|---|
minimax-m3standard · 文字 | opencode-go | $0.30 / $1.20 |
- · 參考連結 · GitHub search · MiniMax · commits · rodrigueztaylor87/rod · @rodrigueztaylor87
- · 參考連結 · GitHub search · MiniMax · commits · SilverKnightKMA/paseo-plugins · @SilverKnightKMA
- · 參考連結 · GitHub search · MiniMax · commits · SilverKnightKMA/paseo-plugins · @SilverKnightKMA
摘錄
now carry the FULL pi reference '<pi-provider>/<model>' (magpie/minimax/MiniMax-M3),
- · 參考連結 · GitHub search · MiniMax · commits · Lore-Hex/quill-router · @jperla
- · 參考連結 · GitHub search · MiniMax · pulls · xiaoY233/Chat2API · @octo-patch
摘錄
feat(minimax): add MiniMax-M3 model entry
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @Vincent-PRO-AI
摘錄
[Model][MiniMax-M3] SM120 Blackwell NVFP4 KV cache Triton reader
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @Pb314314
摘錄
use sparse MLA (dense MLA, DeepSeek-V4 / V4.1, MiniMax-M3, non-MLA models) are not affected; the reason for each is in "Per-model check".
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @cjackal
摘錄
- Pixtral, Cohere Compass and MiniMax-M3 also call `get_cu_seqlens_padding_logics`, so every encoder-CG model whose `cu_seqlens` comes from `maybe_recompute_cu_seqlens` takes its padding from one place. Without
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @MrGeva
摘錄
[Perf][MiniMax-M3] Two-shot fused MoE finalize + all-reduce + norm + MXFP8 up to 512 tokens
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @MrGeva
摘錄
[Perf][MiniMax-M3] Raise the fused MoE finalize + all-reduce cap from 32 to 128 tokens
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @afriedri
摘錄
short prompts at high concurrency is higher on MiniMax-M3 (TP8: up to +21% at 1024 tokens, conc 64; TP4: +14% at 1024
- · 參考連結 · GitHub search · MiniMax · pulls · vllm-project/vllm · @wolfcomos
摘錄
Add `minimax_m3_msa_decode_fp16_dot`, an opt-in optimization for MiniMax-M3 Triton sparse decode on SM100-family GPUs. It uses FP16 dot operands with FP32 accumulation and selects splits targeting
- · 參考連結 · GitHub search · MiniMax · pulls · viztor/dsh-opencode-patch · @viztor
摘錄
| `opencode-go/minimax-m3` | info | `[KNOWN_DEFECT]` toggle-only model offered Minimal…Max | Off / On (D1) |
- · 參考連結 · GitHub search · MiniMax · pulls · thequanhcm/quanpt1 · @thequanhcm
- · 參考連結 · GitHub search · MiniMax · pulls · tenstorrent/tt-metal · @mdragulaTT
- · 參考連結 · GitHub search · MiniMax · pulls · tenstorrent/tt-metal · @mdragulaTT
摘錄
The MiniMax-M3 vision encoder on one Blackhole chip. It turns one image, as the media module's `prepare` resizes it, into its vision tokens: bf16 `[n_tokens,
- · 參考連結 · GitHub search · MiniMax · pulls · sgl-project/sglang · @alphabetc1
- · 參考連結 · GitHub search · MiniMax · pulls · openJiuwen-ai/jiuwenswarm · @octo-patch
- · 參考連結 · GitHub search · MiniMax · pulls · jundot/omlx · @begcdn
摘錄
On M2 Ultra (`applegpu_g14d`) the MiniMax-M3 sparse prefill path crashes once a prompt crosses 4096 tokens: `msa_prefill_supported` gates only on shape (`D=128`, `block_size=128`, `qhead_per_kv=16`), so
- · 參考連結 · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
- · 參考連結 · GitHub search · MiniMax · pulls · inference-sim/inference-sim · @sriumcp
摘錄
the substrate's layer 8 and 9 price changes (MiniMax-M3 routers replicated in fp32, MXFP8 scale bytes, dsv4-nvfp4 MTP experts at MXFP4, a shared lm_head for
- · 參考連結 · GitHub search · MiniMax · pulls · inference-sim/blis-vllm · @sriumcp
摘錄
(`vllm/model_executor/models/config.py:43-55`) for every graph that is not DeepSeek-V4; MiniMax-M3's pool at the same memory is 72645 blocks, not 72893, because the catalog graph now holds
- · 參考連結 · GitHub search · MiniMax · pulls · inference-sim/blis-vllm · @sriumcp
摘錄
the figures on the H200 rows and the MiniMax-M3 points are development-split, in-sample numbers.
- · 參考連結 · GitHub search · MiniMax · pulls · goonercbot/hermes-agent · @goonercbot
摘錄
- Real configured MiniMax-M3 request through corrected Hermes adapter returned exactly `RECOVERY_OK`, end_turn, original request unchanged.
- · 參考連結 · GitHub search · MiniMax · pulls · ghuntley/how-to-build-a-coding-agent · @octo-patch
摘錄
Anthropic-compatible defaults to the standalone workshop agents, with MiniMax-M3.1-Flash as the default and MiniMax-M3 as the alternative.
- · 參考連結 · GitHub search · MiniMax · pulls · boelabs/ai-gateway · @jhisdev
摘錄
- `bun run test` fails: `vercel.test.ts` pins `vercel/minimax/minimax-m3` reasoning as `fixed ["high"]`; the regenerated snapshot now says `openai_effort ["none","high"]`. Tests may not be edited by
- · 參考連結 · GitHub search · MiniMax · pulls · ashhart/TensorFold · @ashhart
摘錄
CUDA decode attention (`window_attention`, shared by Qwen3.5/3.8, Llama, MiniMax-M3, Spark 2.5, Step5 and LFM): `window_rows` merged every chunk's partial for every row, 128 chunks at
- · 參考連結 · GitHub search · MiniMax · pulls · anomalyco/models.dev · @rjuez00
摘錄
- `MiniMax-M3` is no longer listed and a request using that ID returns `model=MiniMax-M3.1-Flash`, so the old entry is removed rather than kept as an
- · 參考連結 · GitHub search · MiniMax · pulls · ROCm/aiter · @zhiding512
摘錄
On gfx1250, every MiniMax-M3 sparse attention call (decode, MTP verify, prefill) runs FlyDSL `pa_decode` (`pa_fwd_asm` does not support gfx1250), and dense layers use it with
- · 參考連結 · GitHub search · MiniMax · pulls · ROCm/aiter · @RElbers
- · 參考連結 · GitHub search · MiniMax · pulls · ROCm/ATOM · @yhl-amd
摘錄
under `multi`, which is the prefill-side setup for MiniMax-M3 PD with UMBP:
- · 參考連結 · GitHub search · MiniMax · pulls · ROCm/ATOM · @rbrugaro-amd
摘錄
does for the target (vLLM already did this). MiniMax-M3's DSpark draft now validates this config instead of building its own, and declares `draft_skips_online_quant` so its
- · 參考連結 · GitHub search · MiniMax · pulls · Quarkgluonmixture/quarkspace · @github-actions[bot]
摘錄
- `intelligence` — minimax-m3 (default) intelligence: catalog 45.4, archive 18.9
- · 參考連結 · GitHub search · MiniMax · pulls · Portkey-AI/models · @polease
- · 參考連結 · GitHub search · MiniMax · pulls · Pinvou/pinvou-agent · @asto18089
- · 參考連結 · GitHub search · MiniMax · pulls · Pb314314/vllm · @Pb314314
摘錄
use sparse MLA (dense MLA, DeepSeek-V4 / V4.1, MiniMax-M3, non-MLA models) are not affected; the reason for each is in "Per-model check".
- · 參考連結 · GitHub search · MiniMax · pulls · NVIDIA/TensorRT-LLM · @ziyixiong-nv
摘錄
sampler, speculative and QSA tests; the MXFP8 and MiniMax-M3 dispatch tests; GSM8K with DSpark (TP4);
- · 參考連結 · GitHub search · MiniMax · pulls · NVIDIA-NeMo/Megatron-Bridge · @pruprakash
摘錄
seeing `TEDotProductAttention`; GREEN 4/4, full file 86 passed, Gemma4/MiniMax-M3/Kimi-K2.5-VL/GPT-OSS unit tests 100 passed. E2E: gpt-oss-20b card inference (TP2 PP2 EP2, `--attention-backend local`) on this branch
- · 參考連結 · GitHub search · MiniMax · pulls · EvolvingLMMs-Lab/lmms-eval · @sahirema
摘錄
sample by frame rate ignore it, e.g. vLLM's MiniMax-M3 and Qwen3-VL loaders. On those, `max_frames_num` has no effect: long clips are sampled at the loader's
- · 參考連結 · GitHub search · MiniMax · pulls · overnightworks/speech · @dependabot[bot]
- · 參考連結 · GitHub search · MiniMax · pulls · maximhq/bifrost · @jimaldon
摘錄
model catalog datasheet already carries `tencent/deepseek-v4-pro`, `tencent/deepseek-v4-flash` and `tencent/minimax-m3`; live-only models are listed with blank pricing and can be priced via overrides.
- · 參考連結 · GitHub search · MiniMax · pulls · maximhq/bifrost · @octo-patch
摘錄
API. New text examples and integration defaults select `MiniMax-M3.1-Flash`, with `MiniMax-M3` as the other current choice. Default management model suggestions put those models first and
- · 參考連結 · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
- · 參考連結 · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
- · 參考連結 · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
- · 參考連結 · GitHub search · MiniMax · pulls · cbc02009/container-images · @mahoro-ai
摘錄
`/v1/chat/completions` request to a messages-protocol Go model (e.g. `opencode_go/minimax-m3`, `qwen3.8-max`, `claude-haiku-5-5`) dies mid-stream with `TypeError: 'NoneType' object is not an iterator` → `APIConnectionError` 500 →
- · 參考連結 · GitHub search · MiniMax · pulls · MrGeva/vllm · @MrGeva
摘錄
320 questions, concurrency 32, same node, together with MiniMax-M3 model changes):
- · 目錄記錄 · GitHub · MiniMax-AI/minimax-code · code
摘錄
afety margin. */ export function computeCompactionTriggerAt(input: { modelId?: string; contextWindow: number; perTurnMaxTokens: number; reserveTokens: number; safetyMarginTokens: number; }): number { if ( input.modelId === 'MiniMax-M3' && (input.contextWindow === 512_000 || input.contextWindow === 1_000_000) ) { return Math.floor(input.contextWindow * 0.9); } const perTurnMaxTokens = Number.isFinite(input.perTurnMaxTokens) && input.perTurnMaxTokens > 0 ? input.perTurnMaxTokens : 0; const reserve = Math.max(input.reserveTokens, perTurnMaxTokens + input.safetyMarginTokens); return Math.max(1, input.contextWindow - reserve); }
此程式碼引用並不能證實這是實際使用中的模型識別碼。
規格
MiniMax M3 是 MiniMax 的 AI 模型，權重開放，於 2026-06-01 發布。 上下文視窗為1,000,000 token。
- 上下文視窗
- 1,000,000
- 輸入
- 影像 · 文字 · 影片
- 輸出
- 文字
- 權重
- 開放
常見問題
MiniMax M3 的上下文視窗有多大？
上下文視窗為1,000,000 token。
MiniMax M3 的權重是否開放？
是的，MiniMax 已公開權重。
MiniMax M3 何時發布？
2026-06-01。