Alex Getmansinyal AI awal
ID

MiniMax M3

MiniMax

Dirilis Diperbarui
Jendela konteks
1M
1.000.000 token
Masukan / Keluaran
gambar, teks, video → teks

Linimasa

  1. · Versi bobot · tautan bukti
Harga masukan / keluaran · USD / 1 juta token
ModelTautan buktiHarga masukan / keluaran
minimax-m3
standard · teks
opencode-go$0,30 / $1,20
  1. · Tautan bukti · GitHub search · MiniMax · commits · rodrigueztaylor87/rod · @rodrigueztaylor87
  2. · Tautan bukti · GitHub search · MiniMax · commits · SilverKnightKMA/paseo-plugins · @SilverKnightKMA
  3. · Tautan bukti · GitHub search · MiniMax · commits · SilverKnightKMA/paseo-plugins · @SilverKnightKMA
    Kutipan
    now carry the FULL pi reference '<pi-provider>/<model>' (magpie/minimax/MiniMax-M3),
  4. · Tautan bukti · GitHub search · MiniMax · commits · Lore-Hex/quill-router · @jperla
  5. · Tautan bukti · GitHub search · MiniMax · pulls · xiaoY233/Chat2API · @octo-patch
    Kutipan
    feat(minimax): add MiniMax-M3 model entry
  6. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @Vincent-PRO-AI
    Kutipan
    [Model][MiniMax-M3] SM120 Blackwell NVFP4 KV cache Triton reader
  7. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @Pb314314
    Kutipan
    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".
  8. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @cjackal
    Kutipan
    - 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
  9. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @MrGeva
    Kutipan
    [Perf][MiniMax-M3] Two-shot fused MoE finalize + all-reduce + norm + MXFP8 up to 512 tokens
  10. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @MrGeva
    Kutipan
    [Perf][MiniMax-M3] Raise the fused MoE finalize + all-reduce cap from 32 to 128 tokens
  11. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @afriedri
    Kutipan
    short prompts at high concurrency is higher on MiniMax-M3 (TP8: up to +21% at 1024 tokens, conc 64; TP4: +14% at 1024
  12. · Tautan bukti · GitHub search · MiniMax · pulls · vllm-project/vllm · @wolfcomos
    Kutipan
    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
  13. · Tautan bukti · GitHub search · MiniMax · pulls · viztor/dsh-opencode-patch · @viztor
    Kutipan
    | `opencode-go/minimax-m3` | info | `[KNOWN_DEFECT]` toggle-only model offered Minimal…Max | Off / On (D1) |
  14. · Tautan bukti · GitHub search · MiniMax · pulls · thequanhcm/quanpt1 · @thequanhcm
  15. · Tautan bukti · GitHub search · MiniMax · pulls · tenstorrent/tt-metal · @mdragulaTT
  16. · Tautan bukti · GitHub search · MiniMax · pulls · tenstorrent/tt-metal · @mdragulaTT
    Kutipan
    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,
  17. · Tautan bukti · GitHub search · MiniMax · pulls · sgl-project/sglang · @alphabetc1
  18. · Tautan bukti · GitHub search · MiniMax · pulls · openJiuwen-ai/jiuwenswarm · @octo-patch
  19. · Tautan bukti · GitHub search · MiniMax · pulls · jundot/omlx · @begcdn
    Kutipan
    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
  20. · Tautan bukti · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
  21. · Tautan bukti · GitHub search · MiniMax · pulls · inference-sim/inference-sim · @sriumcp
    Kutipan
    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
  22. · Tautan bukti · GitHub search · MiniMax · pulls · inference-sim/blis-vllm · @sriumcp
    Kutipan
    (`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
  23. · Tautan bukti · GitHub search · MiniMax · pulls · inference-sim/blis-vllm · @sriumcp
    Kutipan
    the figures on the H200 rows and the MiniMax-M3 points are development-split, in-sample numbers.
  24. · Tautan bukti · GitHub search · MiniMax · pulls · goonercbot/hermes-agent · @goonercbot
    Kutipan
    - Real configured MiniMax-M3 request through corrected Hermes adapter returned exactly `RECOVERY_OK`, end_turn, original request unchanged.
  25. · Tautan bukti · GitHub search · MiniMax · pulls · ghuntley/how-to-build-a-coding-agent · @octo-patch
    Kutipan
    Anthropic-compatible defaults to the standalone workshop agents, with MiniMax-M3.1-Flash as the default and MiniMax-M3 as the alternative.
  26. · Tautan bukti · GitHub search · MiniMax · pulls · boelabs/ai-gateway · @jhisdev
    Kutipan
    - `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
  27. · Tautan bukti · GitHub search · MiniMax · pulls · ashhart/TensorFold · @ashhart
    Kutipan
    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
  28. · Tautan bukti · GitHub search · MiniMax · pulls · anomalyco/models.dev · @rjuez00
    Kutipan
    - `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
  29. · Tautan bukti · GitHub search · MiniMax · pulls · ROCm/aiter · @zhiding512
    Kutipan
    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
  30. · Tautan bukti · GitHub search · MiniMax · pulls · ROCm/aiter · @RElbers
  31. · Tautan bukti · GitHub search · MiniMax · pulls · ROCm/ATOM · @yhl-amd
    Kutipan
    under `multi`, which is the prefill-side setup for MiniMax-M3 PD with UMBP:
  32. · Tautan bukti · GitHub search · MiniMax · pulls · ROCm/ATOM · @rbrugaro-amd
    Kutipan
    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
  33. · Tautan bukti · GitHub search · MiniMax · pulls · Quarkgluonmixture/quarkspace · @github-actions[bot]
    Kutipan
    - `intelligence` — minimax-m3 (default) intelligence: catalog 45.4, archive 18.9
  34. · Tautan bukti · GitHub search · MiniMax · pulls · Portkey-AI/models · @polease
  35. · Tautan bukti · GitHub search · MiniMax · pulls · Pinvou/pinvou-agent · @asto18089
  36. · Tautan bukti · GitHub search · MiniMax · pulls · Pb314314/vllm · @Pb314314
    Kutipan
    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".
  37. · Tautan bukti · GitHub search · MiniMax · pulls · NVIDIA/TensorRT-LLM · @ziyixiong-nv
    Kutipan
    sampler, speculative and QSA tests; the MXFP8 and MiniMax-M3 dispatch tests; GSM8K with DSpark (TP4);
  38. · Tautan bukti · GitHub search · MiniMax · pulls · NVIDIA-NeMo/Megatron-Bridge · @pruprakash
    Kutipan
    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
  39. · Tautan bukti · GitHub search · MiniMax · pulls · EvolvingLMMs-Lab/lmms-eval · @sahirema
    Kutipan
    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
  40. · Tautan bukti · GitHub search · MiniMax · pulls · overnightworks/speech · @dependabot[bot]
  41. · Tautan bukti · GitHub search · MiniMax · pulls · maximhq/bifrost · @jimaldon
    Kutipan
    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.
  42. · Tautan bukti · GitHub search · MiniMax · pulls · maximhq/bifrost · @octo-patch
    Kutipan
    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
  43. · Tautan bukti · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
  44. · Tautan bukti · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
  45. · Tautan bukti · GitHub search · MiniMax · pulls · iv-org/invidious · @Fijxu
  46. · Tautan bukti · GitHub search · MiniMax · pulls · cbc02009/container-images · @mahoro-ai
    Kutipan
    `/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 →
  47. · Tautan bukti · GitHub search · MiniMax · pulls · MrGeva/vllm · @MrGeva
    Kutipan
    320 questions, concurrency 32, same node, together with MiniMax-M3 model changes):
  48. · Entri katalog · GitHub · MiniMax-AI/minimax-code · code
    Kutipan
    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); }

    Referensi dalam kode ini tidak mengonfirmasi pengenal model yang digunakan.

Spesifikasi

MiniMax M3 adalah model AI dari MiniMax dengan bobot terbuka, dirilis pada 2026-06-01. Jendela konteksnya sebesar 1.000.000 token.

Jendela konteks
1.000.000
Masukan
gambar · teks · video
Keluaran
teks
Bobot
Terbuka

Pertanyaan

Seberapa besar jendela konteks MiniMax M3?

Jendela konteksnya sebesar 1.000.000 token.

Apakah bobot model MiniMax M3 tersedia untuk umum?

Ya. MiniMax menerbitkan bobot model.

Kapan MiniMax M3 dirilis?

2026-06-01.