Mixedbread: $5 one-time free credits, no credit card
Source: https://www.mixedbread.com/pricing
Description
Create account to comment on specific lines or Sign in
+
1
Mixedbread (Berlin-based AI search/retrieval platform — best known for the open-source mxbai-embed-large and mxbai-rerank model families) offers a Starter tier with $5 one-time credits and no credit card required to sign up. The credits are spent against usage-based pricing for embeddings (indexing), search queries, reranking, and storage on their hosted Stores (managed vector index + retrieval pipeline). Free tier is rate-limited to 100 requests/min and capped at 10 stores + 3 workspace users.
No comments on this line yet.
+ 2
No comments on this line yet.
+ 3 This is a true "try before you buy" play — small enough that you'll burn through it on a serious indexing job, but generous enough to validate semantic search / RAG / multimodal retrieval against a real index without committing a card.
No comments on this line yet.
+ 4
No comments on this line yet.
+
5
No comments on this line yet.
+ 6
No comments on this line yet.
+ 8
No comments on this line yet.
+ 9 1. Go to mixedbread.com and click "Sign in" (top-right) — or jump straight to the pricing page and click "Start free" on the Starter card.
No comments on this line yet.
+ 10 2. Sign up with email or Google / GitHub OAuth — no credit card prompted.
No comments on this line yet.
+ 11 3. Confirm your email if prompted.
No comments on this line yet.
+ 12 4. You land in the Mixedbread Platform dashboard. The $5 credit is auto-applied to the workspace.
No comments on this line yet.
+ 13 5. Create an API key under Settings → API Keys.
No comments on this line yet.
+ 14 6. Create your first Store (managed retrieval index) via the dashboard or the Quickstart — or hit the inference endpoints directly with the Python / Node SDK.
No comments on this line yet.
+ 15
No comments on this line yet.
+ 16 Important:
No comments on this line yet.
+ 17 • No credit card on file — when the $5 runs out, the API just stops serving 402 errors. Nothing auto-bills.
No comments on this line yet.
+ 18 • Credits are one-time, per workspace (don't expect to spin up new workspaces to refresh them).
No comments on this line yet.
+ 19 • The Starter tier is for experimentation — for anything production you'll want Scale ($20/mo flat + usage).
No comments on this line yet.
+ 20
No comments on this line yet.
+
21
No comments on this line yet.
+ 22
No comments on this line yet.
+ 24
No comments on this line yet.
+ 25 The credits draw down against Mixedbread's usage-based pricing:
No comments on this line yet.
+ 26
No comments on this line yet.
+ 27 ServicePriceWhat $5 buys
No comments on this line yet.
+ 28 Indexing — Fast (embed for ingestion)$1.50 / 1M tokens~3.3M tokens of content indexed
No comments on this line yet.
+ 29 Indexing — High Quality$3 / 1M tokens~1.6M tokens indexed
No comments on this line yet.
+ 30 Search — Basic (vector retrieval)$4 / 1K queries~1,250 queries
No comments on this line yet.
+ 31 Search — Rerank (vector + cross-encoder rerank)$7.50 / 1K queries~660 queries
No comments on this line yet.
+ 32 Storage$0.50 / 1M tokens / month10M tokens stored for 1 month
No comments on this line yet.
+ 33
No comments on this line yet.
+ 34 In practice: you can index a small-to-medium knowledge base (tens of thousands of chunks), run a few hundred reranked queries, and host it for a month — all on the free $5.
No comments on this line yet.
+ 35
No comments on this line yet.
+
36
No comments on this line yet.
+ 37
No comments on this line yet.
+ 39
No comments on this line yet.
+ 40 All models run on Mixedbread's hosted endpoints — no GPU needed locally.
No comments on this line yet.
+ 41
No comments on this line yet.
+
43
• mxbai-embed-large-v1 — flagship English embedding model, 335M params, 1024 dims, 512-token context. SOTA among BERT-large-sized models on MTEB at release; outperforms OpenAI text-embedding-3-large on several benchmarks.
No comments on this line yet.
+
44
• mxbai-embed-xsmall-v1 — efficient retrieval-focused embedding, 22.7M params, 384 dims. Built for high-throughput / low-latency.
No comments on this line yet.
+ 45 • ColBERT (multi-vector / late interaction) — contextualized late-interaction retrieval; combines dense vectors with cross-encoder-style scoring.
No comments on this line yet.
+ 46
No comments on this line yet.
+
48
• mxbai-rerank-large-v2 — flagship reranker, multilingual (100+ languages), SOTA accuracy.
No comments on this line yet.
+
49
• mxbai-rerank-base-v2 — smaller / faster reranker, also multilingual.
No comments on this line yet.
+
50
• mxbai-rerank-large-v1 / mxbai-rerank-base-v1 — first-generation, English-focused.
No comments on this line yet.
+ 51
No comments on this line yet.
+ 53 • Document Parsing — managed OCR + structure extraction (PDFs, docs, images) feeding directly into Stores.
No comments on this line yet.
+ 54 • Stores — end-to-end managed retrieval: chunking, embedding, indexing, hybrid search, rerank, MCP server access.
No comments on this line yet.
+ 55
No comments on this line yet.
+ 56 Source: Mixedbread Models Docs · Embeddings Inference Docs
No comments on this line yet.
+ 57
No comments on this line yet.
+
58
No comments on this line yet.
+ 59
No comments on this line yet.
+ 61
No comments on this line yet.
+ 62 LimitStarter (Free)Scale ($20/mo)Enterprise
No comments on this line yet.
+ 63 Credits included$5 one-time$20/mo included + usageCustom
No comments on this line yet.
+ 64 Query rate limit100 requests / min1,200 queries/min, 360 ingestion/minCustom
No comments on this line yet.
+ 65 Stores1010,000Unlimited
No comments on this line yet.
+ 66 Workspace users3UnlimitedUnlimited
No comments on this line yet.
+ 67 SupportCommunity SlackPriority Slack (same-day SLA)Dedicated team
No comments on this line yet.
+ 68 Backups—AutomaticAuto + point-in-time recovery
No comments on this line yet.
+ 69 BYOC——Yes
No comments on this line yet.
+ 70
No comments on this line yet.
+ 71 The 1,000 store-files number sometimes quoted in third-party trackers refers to indexed file count headroom on the Starter tier — confirm in-dashboard, since limits get tuned regularly.
No comments on this line yet.
+ 72
No comments on this line yet.
+
73
No comments on this line yet.
+ 74
No comments on this line yet.
+ 76
No comments on this line yet.
+ 77 • No credit card means hard stop, not auto-charge — great for experimentation, but plan to upgrade before a launch.
No comments on this line yet.
+
78
• Use xsmall for indexing huge corpora — at 384 dims and lower per-token cost, you'll get more bang for the $5 than large-v1.
No comments on this line yet.
+ 79 • Rerank only what matters — at $7.50/1K queries, restrict reranking to top-K candidates from cheap vector search.
No comments on this line yet.
+
80
• Open-source escape hatch — if you outgrow the credit, every flagship model (mxbai-embed-large-v1, mxbai-rerank-large-v2, etc.) is on Hugging Face under permissive licenses; you can self-host on any GPU box (or even CPU for the small variants).
No comments on this line yet.
+ 81 • Startup program — qualified startups can apply for $250 in one-time credits (ask via the platform / their team).
No comments on this line yet.
+ 82 • Region — hosted globally; data residency / EU-only deployment available on Enterprise (BYOC).
No comments on this line yet.
+ 83
No comments on this line yet.
+
84
No comments on this line yet.
+ 85
No comments on this line yet.
+ 86 Sources:
No comments on this line yet.
+ 87 • Mixedbread Pricing
No comments on this line yet.
+ 88 • Mixedbread Docs Hub
No comments on this line yet.
+ 89 • Embeddings Inference Docs
No comments on this line yet.
+ 90 • mxbai-embed-large-v1 Model Card
No comments on this line yet.
+ 91 • mxbai-rerank-large-v2 Model Page
No comments on this line yet.
+ 92 • mxbai-rerank GitHub
No comments on this line yet.
+ 93 • Hugging Face: mixedbread-ai org
No comments on this line yet.