# SiliconFlow

> SiliconFlow is a hosted inference API for open-weight models, covering chat, vision, embeddings, reranking, image, video and speech. It answers OpenAI-style and Anthropic-style calls at api.siliconflow.com with a Bearer key.

- Canonical: https://www.anchorterminal.com/tools/siliconflow
- Markdown: https://www.anchorterminal.com/tools/siliconflow.md (~9,150 tokens)
- Slim: https://www.anchorterminal.com/tools/siliconflow.min.md (~2,130 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/tools/siliconflow.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- Updated: 2026-10-10

## Overview

**Grade D · 46.7/100 · rank #848 of 950 · #17 in Model APIs & inference · not agent-ready · confidence medium**


## Assessment

Per-token prices for every listed model are public, and one key reaches chat, embeddings, reranking, image, video and speech through OpenAI-style and Anthropic-style routes. No status page, SLA, security page or official SDK was found, release notes stop at 11 June 2026, and several model removals are dated the same day as their notice.

## Facts

| Field | Value |
| --- | --- |
| Vendor | SiliconFlow Labs Pte. Ltd. (https://www.siliconflow.com) |
| Kind | Model API |
| Category | Model APIs & inference (https://www.anchorterminal.com/categories/inference) |
| Transport | HTTP |
| Endpoint | `https://api.siliconflow.com/v1` |
| Auth | API key · Self-serve Bearer key from the console at https://cloud.siliconflow.com/account/ak after a browser sign-up with a Google or GitHub account. The Anthropic-style `/messages` route also takes the key in `x-api-key`. No scopes, expiry or per-key limits are documented, and rate limits apply to the account, not the key. The error page says some models answer 403 until real-name authentication is done. |
| Pricing | Pay per use (from $0.15 / 1M in) · Pay per token, image, video or byte of speech text, with $1 in free credits for a new account. Whether a card is needed to claim the credit was not established. DeepSeek-V4.1-Flash costs $0.15 in, $0.003 cached and $0.60 out per 1M tokens. The FAQ says monthly spending limits can be set in the dashboard (https://www.siliconflow.com/pricing, checked 2026-10-09). |
| x402 | No · No x402, MPP or L402 in the docs index, the OpenAPI file or the pricing page (checked 2026-10-09). |
| Licence | Proprietary service under the SiliconFlow Terms of Use |
| Docs | https://docs.siliconflow.com/en/userguide/quickstart |
| llms.txt | https://docs.siliconflow.com/llms.txt |
| Last release | 2026-09-14 |
| Endpoints | 22 operations under https://api.siliconflow.com/v1 in the OpenAPI file. `/chat/completions`, `/completions`, `/messages` (Anthropic-style), `/embeddings`, `/rerank`, `/images/generations`, `/video/submit` and `/video/status`, `/audio/speech`, `/audio/transcriptions`, voice upload and list, `/files`, `/batches`, `/models`, `/user/info` and `/systemone` (alpha) |
| Models on 9 October 2026 | The pricing page lists chat models from DeepSeek, Qwen, Z.ai, Moonshot AI, MiniMax, Tencent, Google (Gemma) and OpenAI (gpt-oss), FLUX and Z-Image for images, Wan2.2 for video and three text-to-speech models. A full count was not taken, because the tables load more rows on request |
| Rate limits | Per account and per model, set by monthly spend tier. L0 1,000 requests and 40,000 tokens a minute, L1 1,200 and 60,000, L2 2,000 and 80,000, L3 4,000 and 160,000, L4 8,000 and 500,000, L5 10,000 and 2,000,000. DeepSeek-R1 and DeepSeek-V3 also have 30 requests an hour. Per-model figures are in the console |
| Errors | JSON body with a numeric `code`, a `message` and `data`, for example `20012` for an unknown model. 400, 401, 403, 429, 500, 503 and 504 are described. The 429 message names the limit reached, such as TPM. No `Retry-After` header is documented |
| Tool calling | OpenAI-style `tools` with up to 128 functions. Each model page says whether Tools are supported. No `tool_choice` field is in the OpenAPI file's chat schema |
| Structured output | `response_format` of `json_object`. The DeepSeek-V4.1-Flash model page says JSON Mode is supported and Structured Outputs are not |
| Prompt caching | Cached input prices are published per model, such as $0.003 per 1M tokens on DeepSeek-V4.1-Flash. No caching guide is in the docs index |
| Batch | `/files` and `/batches` in the OpenAPI file, for `/v1/chat/completions` only. The field description gives a maximum window of 24 hours and a minimum of 336 hours |
| Credentials | API keys created in the console at cloud.siliconflow.com/account/ak, sent as a Bearer token, or in `x-api-key` on `/messages`. No scopes, expiry or per-key limits are documented. Sign-up is through a Google or GitHub account |
| Deprecation notices | Dated entries in the release notes. Ten model removal notices between 31 December 2025 and 11 June 2026, six of them dated the day of removal. No stated notice period |
| Data use in the terms | Clause 7.4 says SiliconFlow will not access Interaction Data, has no obligation to store it, will not use or disclose it without authorisation, and uses it only to supply the service. No retention period and no statement on model training were found |
| Generated media | Image and video URLs are valid for one hour, per the API reference |
| Contract | SiliconFlow Labs Pte. Ltd., Singapore law, SIAC arbitration in Singapore. Fees are not refunded once due, and a fee dispute must be raised within seven days |
| Support | help@siliconflow.com and contact@siliconflow.com, and a Discord server |
| Capabilities | inference.llm, inference.open-weights, embed.text, rerank, image.generate, video.generate, speech.tts, speech.stt |
| Tags | hosted, model, open-weights, usage-based, free-credit, openapi, llms-txt, openai-compatible, anthropic-compatible, singapore |
| JSON | https://www.anchorterminal.com/api/v1/tools/siliconflow.json |

## Score breakdown (methodology v0.4, October 2026 research run)

Assessed 2026-10-09 from public evidence against the published checklist (https://www.anchorterminal.com/benchmark/#checklist). Confidence: medium. Performance and Task success pending (no score, not in the total); the total is Σ(score × weight) ÷ 80 over the 7 assessed categories. "This run" is each category's share of the 100 points.

| Category | Weight | This run | Score (0–100) | Points |
| --- | --- | --- | --- | --- |
| Reliability | 16% | 20 | 33 | 6.6 |
| Performance | 10% | pending | pending | n/a |
| Schema & documentation | 13% | 16.2 | 65 | 10.6 |
| Agent ergonomics | 13% | 16.2 | 60 | 9.8 |
| Security & auth | 14% | 17.5 | 39 | 6.8 |
| Payments & pricing | 10% | 12.5 | 35 | 4.4 |
| Task success | 10% | pending | pending | n/a |
| Maintenance & community | 7% | 8.8 | 47 | 4.1 |
| Transparency & trust (editorial 34, provenance 68) | 7% | 8.8 | 51 | 4.5 |
| Negative events | up to −15 | up to −15 | none recorded | 0 |
| **Total** | | | | **46.7 → D** |

### Why each score

- Reliability 33: Hosted reading. No status page is linked from the site, the docs index or any page read (0). With no page there is no incident record to read (5 of 30). Rate limits are published with numbers by spend tier, from 1,000 requests and 40,000 tokens a minute at L0 to 10,000 and 2,000,000 at L5, on a page served only in Chinese, with per-model figures left to the console (12 of 15). The 429 body is documented and names the limit reached, and the docs say to wait and retry. No `Retry-After` header, backoff figures or idempotency guidance were found (6 of 15). No SLA found. Clause 4.2 of the terms disclaims any uptime or availability commitment (0). The serverless API is generally available, and `/systemone` is marked alpha (10).
- Performance: Pending. Latency is measured per call by our probes, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until the first probe window closes.
- Schema & documentation 65: Model reading. Public OpenAPI 3.0 file with 22 operations and 102 schemas, linked from `llms.txt` under the Chinese path only. Its chat model enum of 71 ids includes removed models and omits every model added since July 2026 (19 of 25). `llms.txt` and a Markdown copy of each docs page (10). Guides for function calling, JSON mode, reasoning, prefix and FIM completion. The function calling guide lists only models the release notes show as removed, and its examples use DeepSeek-V2.5 (10 of 20). Typed parameters with 104 enums and required fields. `response_format` is typed as a bare string with an example (10 of 15). Curl and Python examples, an error page with seven HTTP codes and one body code, and 400, 401, 404, 429, 503 and 504 responses in the file (9 of 15). The file's version reads 1.0.0. The release notes are dated, stop at 11 June 2026 and since March 2026 record only model removals (7 of 15).
- Agent ergonomics 60: Model reading of the checklist (tool use, structured output, caching, context, batch, SDKs, errors). OpenAI-style `tools` with up to 128 functions, and each model page says whether Tools are supported. No `tool_choice` field is in the chat schema and the guide is out of date (12 of 20). `response_format` of `json_object`. The DeepSeek-V4.1-Flash page says Structured Outputs are not supported (7 of 15). Cached input prices are published per model. No caching guide was found (8 of 15). Context listed as 1049K tokens on eleven chat models on the home page, with output up to 393K (14 of 15). `/files` and `/batches` are in the OpenAPI file for chat completions only, with no guide in the English index, no batch price and a window description that contradicts itself (5 of 10). No official SDK. The OpenAI and Anthropic clients work with a changed base URL (6 of 10). Errors carry an HTTP status, a numeric code and a message. No list of codes was found (8 of 15).
- Security & auth 39: Model reading. Bearer keys created in the console. No scopes, expiry, per-key limits or rotation are documented, and deletion of a key was not confirmed because the console was not read (15 of 30). The terms say Interaction Data is used only to supply the service and is not used or disclosed without authorisation. No statement on model training was found (12 of 20). The terms say there is no obligation to store Interaction Data. No retention period or zero-retention option is stated, and clause 5.11 allows communications to be recorded and passed to the authorities (6 of 15). The release note of 26 March 2026 describes usage and cost breakdowns and invoice export, and the pricing FAQ says monthly spending limits can be set. No audit log was found (6 of 15). No security.txt (404), disclosure policy, bug bounty or named certification was found on the pages read. The product introduction says the service complies with industry standards and names none (0 of 20).
- Payments & pricing 35: No machine payment protocol (0). Per-token, per-image and per-video prices on the pricing page, read without a login, such as DeepSeek-V4.1-Flash at $0.15 in and $0.60 out per 1M tokens (20). $1 in free credits for a new account. Whether a card is needed was not established, so this line takes 15 of 20. A person signs up in a browser with a Google or GitHub account, and no key API was found (0).
- Task success: Pending. Task success needs the category task suites run through each tool, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until then. A data provider's data-quality score is published on its listing now and becomes half of this category when it's scored.
- Maintenance & community 47: Model reading. Hy4-preview is listed as released on 14 September 2026 and the blog's newest posts are dated 28 September 2026 (30). No notice period is stated. Six removal notices from 31 December 2025 to 11 June 2026 are dated the day of removal, and two in March 2026 gave six and seven days (2 of 12). The release notes record about 60 model ids removed in ten notices over the last twelve months (2 of 8). The release notes stop at 11 June 2026 while nine chat models were added afterwards. Support is by email and Discord (6 of 15). Replies in the support channels were not examined (2 of 10). No official SDK. The `langchain-siliconflow` package in the vendor's GitHub organisation was last pushed on 12 December 2025 (3 of 15). The API description and two guides are out of date (2 of 10).
- Transparency & trust 51: Closed service. The Terms of Use name the entity, Singapore law and SIAC arbitration. They also grant a licence "for in Singapore" and forbid use for any commercial purposes, which sits oddly with a paid API (10 of 30). The terms and the privacy policy agree that Interaction Data is processed on the customer's instructions with no obligation to store it. No retention period or DPA was found, and the privacy policy keeps drafting brackets around several values (12 of 30). Dated removal notices in the release notes, with no written policy or notice period (8 of 20). Stripe is named as payment processor and other providers only by type. Data may be handled within or outside Singapore, with no locations given (4 of 20).

Fix list for a coding agent, everything this grade says the listing lacks, the biggest gain first (26 items): https://www.anchorterminal.com/fixes/siliconflow.md (JSON https://www.anchorterminal.com/fixes/siliconflow.json)

### What we couldn't check

- unchecked: the console at cloud.siliconflow.com, so key deletion, per-key settings, spending limits, per-model rate limits, usage logs and whether a card is needed for the free credit were not confirmed
- unchecked: no request was sent to api.siliconflow.com, so the live model list, the 429 body and headers, and the routing of retired model ids rest on the docs
- unchecked: whether a status page exists. None is linked from the site or the docs pages read, and no address was guessed
- unchecked: the rendered API reference pages. The OpenAPI file was read in their place
- unchecked: the Chinese-language docs beyond the rate limit page, the blog posts themselves, and replies in the Discord server
- unchecked: model ids for GLM-5.3 and Kimi-K3 follow the vendor's naming pattern and were not read from their model pages
- The release notes do not say when each entry was published. Six entries carry a label equal to the removal date. Whether customers were told earlier by email was not established, so no deduction is taken
- The Terms of Use forbid use for any commercial purposes, scraping tools and benchmarking (clauses 3.4 p, i and l). Recorded as a fact with no deduction. It matters before any probe is run
- The lead was right about the interface and the host. The fine-tuning product is console-only and was removed from batch 4. It is not graded here
- The pricing page's audio table is headed per 1M UTF-8 bytes while its footnote says per minute and per 1,000 characters
- `lastRelease` is the release date the site gives for Hy4-preview, the newest model, since the release notes stop at 11 June 2026
- One slip. The RDAP lookup at rdap.org redirected to rdap.verisign.com before that host's robots.txt was read. rdap.org's own robots.txt request answered 400, and api.github.com's answered 404
- No SLA, DPA, sub-processor list, security page or official SDK was found. Whether they exist on request is not established
- Reserved GPUs and fine-tuning share the account and terms and were not graded. The products page marks dedicated endpoints as coming soon

### Sources

- home page with model list, release dates and prices: <https://www.siliconflow.com/> (seen 2026-10-09)
- pricing page: <https://www.siliconflow.com/pricing> (seen 2026-10-09)
- products page: <https://www.siliconflow.com/products> (seen 2026-10-09)
- about page: <https://www.siliconflow.com/about> (seen 2026-10-09)
- blog index: <https://www.siliconflow.com/blog> (seen 2026-10-09)
- DeepSeek-V4.1-Flash model page: <https://www.siliconflow.com/models/deepseek-v4-1-flash> (seen 2026-10-09)
- security.txt (404): <https://www.siliconflow.com/.well-known/security.txt> (seen 2026-10-09)
- robots.txt, site (allows all): <https://www.siliconflow.com/robots.txt> (seen 2026-10-09)
- robots.txt, docs (ai-input=yes): <https://docs.siliconflow.com/robots.txt> (seen 2026-10-09)
- docs index: <https://docs.siliconflow.com/llms.txt> (seen 2026-10-09)
- quick start: <https://docs.siliconflow.com/en/userguide/quickstart.md> (seen 2026-10-09)
- product introduction: <https://docs.siliconflow.com/en/userguide/introduction.md> (seen 2026-10-09)
- error handling: <https://docs.siliconflow.com/en/faqs/error-code.md> (seen 2026-10-09)
- other issues FAQ: <https://docs.siliconflow.com/en/faqs/misc.md> (seen 2026-10-09)
- rate limits (Chinese page, reached from the English error page's link): <https://docs.siliconflow.com/cn/userguide/rate-limits/rate-limit-and-upgradation> (seen 2026-10-09)
- function calling guide: <https://docs.siliconflow.com/en/userguide/guides/function-calling.md> (seen 2026-10-09)
- JSON mode guide: <https://docs.siliconflow.com/en/userguide/guides/json-mode.md> (seen 2026-10-09)
- Claude Code guide: <https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-ClaudeCode.md> (seen 2026-10-09)
- OpenAPI file, read in place of the rendered API reference pages: <https://docs.siliconflow.com/cn/api-reference/openapi.yaml> (seen 2026-10-09)
- release notes: <https://docs.siliconflow.com/en/release-notes/overview.md> (seen 2026-10-09)
- Terms of Use, updated 9 September 2026: <https://docs.siliconflow.com/en/legals/terms-of-service.md> (seen 2026-10-09)
- Privacy Policy, updated 9 September 2026: <https://docs.siliconflow.com/en/legals/privacy-policy.md> (seen 2026-10-09)
- GitHub organisation repositories: <https://api.github.com/orgs/siliconflow/repos?per_page=100&sort=pushed> (seen 2026-10-09)
- GitHub organisation named in the OpenAPI file: <https://api.github.com/orgs/siliconflow-inc/repos?per_page=100&sort=pushed> (seen 2026-10-09)
- domain registration: <https://rdap.org/domain/siliconflow.com> (seen 2026-10-09)

## Who's behind it (provenance 68/100, checked 2026-10-09)

| Check | Finding | Points |
| --- | --- | --- |
| Legal entity named | SiliconFlow Labs Pte. Ltd. | 20/20 |
| Domain age | siliconflow.com, registered 2019-02-14 (7 years) | 11/15 |
| Endpoint on the vendor's domain | api.siliconflow.com | 15/15 |
| Terms of service | read, states 6 of the 7 things a reader expects, and has 3 clauses that cost points | 3.1/10 |
| Privacy policy | read, states 7 of the 8 things a reader expects | 9.3/10 |
| Status page | not found | 0/10 |
| Changelog | published | 10/10 |
| security.txt | not found | 0/10 |

The Terms of Use, updated 9 September 2026, name SILICONFLOW LABS PTE. LTD., apply Singapore law and set SIAC arbitration in Singapore. They govern the platform in its web, API and mobile versions, and no separate API agreement was found.

The Privacy Policy, updated 9 September 2026, names the same entity and cites Singapore's Personal Data Protection Act 2012. It keeps square brackets around several values, such as [Stripe] and [Google and GitHub].

No status page is linked from the site, the docs index or the pages read.

https://www.siliconflow.com/.well-known/security.txt answered 404.

RDAP for siliconflow.com gives a registration date of 2019-02-14.

The API answers at api.siliconflow.com, the docs at docs.siliconflow.com and the console at cloud.siliconflow.com. The terms point users in mainland China to siliconflow.cn.

No SLA, DPA or sub-processor list was found.

### Terms and privacy, as read

A reading by a fixed set of rules, each answered with the vendor's own sentence. Not legal advice.

**Terms of service** (https://docs.siliconflow.com/en/legals/terms-of-service), read 2026-10-09, gives no date, states 6 of the 7 things a reader expects.

- To know. Restricts automated access (costs points). "use any manual or automated software, devices or other processes (including but not limited to spiders, robots, scrapers, crawlers, avatars, data mining tools or similar tools) to “scrape”, collect or download any information and data from the Platform or Services;"
- To know. Restricts benchmarking or competitive use (costs points). "access the Platform or Services in order to build a competitive product or service or otherwise to compete with SiliconFlow;"
- To know. Says the terms or the service can change without notice (costs points). "SiliconFlow may at any time, and without giving any reason or notice, Update the Platform and/or Services without any liability towards you."
- To know. Says access can be ended without notice or for any reason. "…to require you to provide additional Registration Data, change your password, temporarily or permanently suspend or terminate your Account, or impose limits on or restrict your access to and use of the Platform or Services with or without notice at any time for any or no reason including:"
- To know. Requires arbitration or waives class actions. "shall be referred to and finally resolved by arbitration in Singapore administered by the Singapore International Arbitration Centre"
- Not found in the text. Gives the date it was last updated.
- Names the governing law or courts. The law of the Republic of Singapore.
- Says how changes to the terms are announced. Says it gives notice of a change.
- Also in the text (2026-10-08). The terms forbid using the Platform or Services for any commercial purposes, including advertising or selling goods or services. "use the Platform or Services for any commercial purposes, including but not limited to advertising, selling or offering to sell any goods or services, or conducting any trainings, events or courses."
- Also in the text (2026-10-08). SiliconFlow may restrict a user's use of model output at any time and require the user to stop using it and delete copies. "If, at our discretion, we consider that your use of the output violates laws and regulations or may infringe the rights of any third party, we may restrict your use of the output at any time and require you to cease using the output (and delete any copies thereof)"
- Also in the text (2026-10-08). A user who wishes to dispute a fee must notify SiliconFlow in writing within seven days of the charge. "If you dispute or disagree with any Fees charged, you must notify SiliconFlow in writing within seven (7) days from the date such Fees are charged to you."

**Privacy policy** (https://docs.siliconflow.com/en/legals/privacy-policy), read 2026-10-09, gives no date, states 7 of the 8 things a reader expects.

- Not found in the text. Gives the date it was last updated.
- Says how long data is kept. For as long as needed, with no period named.
- Gives a privacy contact. contact@siliconflow.com.
- Also in the text (2026-10-08). SiliconFlow states it has no obligation to store Interaction Data unless laws or the service rules say otherwise. "You understand and further agree that, unless otherwise provided by laws and regulations or as agreed in the service rules, we have no obligation to store your Interaction Data, and we will not assume any responsibility for your data storage efforts or outcomes."

## Live (updated 2026-10-10 02:07 UTC)

- Right now: up, HTTP 404, 261 ms, checked 2026-10-10 02:07 UTC (get on `https://api.siliconflow.com/v1`)
- Uptime 24h 100.0% (107 probes) · 30 days 100.0% (107 probes) · p50 271 ms · p95 383 ms
- Watching deprecations <https://docs.siliconflow.com/en/release-notes/overview>
- Watching pricing <https://www.siliconflow.com/pricing>
- Watching privacy <https://docs.siliconflow.com/en/legals/privacy-policy>
- Watching terms <https://docs.siliconflow.com/en/legals/terms-of-service>
- Always current: https://www.anchorterminal.com/api/v1/live/siliconflow.json

## Probe metrics

Not measured yet. Our benchmark probes haven't run, so there's no availability, latency or error rate from a run and Performance is pending. Live uptime, where we poll the endpoint, is under Live and doesn't change the score.

## Models and prices (per 1M tokens)

| Model | Input | Output | Context | Role | Supports |
| --- | --- | --- | --- | --- | --- |
| `deepseek-ai/DeepSeek-V4.1-Flash` DeepSeek V4.1 Flash | $0.15 | $0.60 | 1.05M | default (context listed as 1049K, cached input $0.003, image input, tools and JSON mode per the model page) | not checked |
 Rate limits depend on your account tier: https://docs.siliconflow.com/cn/userguide/rate-limits/rate-limit-and-upgradation

## Prices

| Item | Price | Unit | Note |
| --- | --- | --- | --- |
| Z-Image-Turbo, one image | $0.005 | per image |  |
| FLUX.1-dev, one image | $0.014 | per image |  |
| FLUX 1.1 [pro], one image | $0.04 | per image |  |
| Wan2.2-T2V-A14B, one video | $0.29 | per call | Priced per video generated |

Across all listings: https://www.anchorterminal.com/prices/index.md

## Dated changes

- 2026-04-29 · Shutdown · `Qwen/QwQ-32B`, `IndexTeam/IndexTTS-2` and six other models removed (source: <https://docs.siliconflow.com/en/release-notes/overview>)
- 2026-05-15 · Shutdown · `moonshotai/Kimi-K2-Instruct`, `zai-org/GLM-4.6` and five other models removed (source: <https://docs.siliconflow.com/en/release-notes/overview>)
- 2026-06-08 · Shutdown · `nex-agi/DeepSeek-V3.1-Nex-N1` removed (source: <https://docs.siliconflow.com/en/release-notes/overview>)
- 2026-06-11 · Shutdown · `zai-org/GLM-4.7` removed, and traffic for `zai-org/GLM-5` and `moonshotai/Kimi-K2.5` routed to GLM-5.1 and Kimi-K2.6 (source: <https://docs.siliconflow.com/en/release-notes/overview>)

All listings, as a calendar: https://www.anchorterminal.com/sunsets.ics

## Strengths

- Per-token, per-image and per-video prices are public for every listed model, such as DeepSeek-V4.1-Flash at $0.15 in and $0.60 out per 1M tokens
- One Bearer key reaches 22 operations in a public OpenAPI 3.0 file, with OpenAI-style `/chat/completions` and Anthropic-style `/messages`
- `llms.txt` and a Markdown copy of every docs page, in English and Chinese
- Rate limits are published by spend tier, from 1,000 requests and 40,000 tokens a minute at L0 to 10,000 and 2,000,000 at L5
- The terms say SiliconFlow will not access Interaction Data, has no obligation to store it, and uses it only to supply the service

## Weaknesses

- No status page, incident history or SLA was found, and the terms disclaim any uptime or availability commitment
- Release notes stop at 11 June 2026. Six notices from December 2025 to June 2026 give a removal date equal to the notice date
- The Terms of Use of 9 September 2026 forbid use for any commercial purposes, scraping tools and benchmarking. This matters before any probe is run
- The OpenAPI model list and the function calling guide name models already removed, and omit every model added since July 2026
- No security.txt, disclosure policy, certification, DPA or sub-processor list was found, and no official SDK is published

## Before you call it (notes for agents)

1. Set the OpenAI client's base URL to `https://api.siliconflow.com/v1`, or the Anthropic client's to `https://api.siliconflow.com/`, and send the key as a Bearer token
2. Take model ids from the model pages or `GET /v1/models`, not from the OpenAPI enum or the function calling guide, which list removed models
3. Check the `model` field of each response. On 11 June 2026 traffic for GLM-5 and Kimi-K2.5 was routed to successor models
4. Use `response_format` of `json_object` only where the model page says JSON Mode is supported, and keep `max_tokens` about 10,000 below the context length
5. Set the Claude Code environment variables by hand. The automated route pipes a script from an Amazon S3 bucket into bash

## Connect

Install:

```bash
pip install --upgrade openai
```

First request:

```bash
curl --request POST \
  --url https://api.siliconflow.com/v1/chat/completions \
  --header 'authorization: Bearer $SILICONFLOW_API_KEY' \
  --header 'content-type: application/json' \
  --data '{"model":"deepseek-ai/DeepSeek-V4.1-Flash","messages":[{"role":"user","content":"Hello"}],"max_tokens":128}'
```

Claude Code:

```bash
export ANTHROPIC_BASE_URL="https://api.siliconflow.com/"
export ANTHROPIC_MODEL="your-preferred-model"
export ANTHROPIC_API_KEY="sk-your-api-key-here"
```

## Similar tools

Ranked by shared capabilities, then score. Same-category tools with no shared capability key are listed last.

| Tool | Grade | Score | Rank | Shared capabilities | x402 | Markdown |
| --- | --- | --- | --- | --- | --- | --- |
| Cloudflare Workers AI | B | 68.3 | 224 | inference.llm, inference.open-weights, embed.text, rerank, image.generate, speech.stt, speech.tts | no | https://www.anchorterminal.com/tools/cloudflare-workers-ai.md |
| LocalAI | B | 68 | 232 | inference.open-weights, embed.text, rerank, speech.stt, speech.tts, image.generate, video.generate | no | https://www.anchorterminal.com/tools/localai.md |
| DeepInfra | B | 63 | 411 | inference.llm, inference.open-weights, embed.text, rerank, image.generate, speech.stt, speech.tts | no | https://www.anchorterminal.com/tools/deepinfra.md |
| Novita AI | D | 53.5 | 706 | inference.llm, inference.open-weights, embed.text, rerank, image.generate, video.generate, speech.tts | no | https://www.anchorterminal.com/tools/novita-ai.md |
| Lemonade | B | 63.8 | 372 | inference.open-weights, embed.text, rerank, speech.stt, speech.tts, image.generate | no | https://www.anchorterminal.com/tools/lemonade.md |
| KoboldCpp | C | 60.5 | 510 | inference.open-weights, embed.text, image.generate, speech.stt, speech.tts | no | https://www.anchorterminal.com/tools/koboldcpp.md |

## Panel reviews (0)

Reviewed by the Anchor panel (https://www.anchorterminal.com/reviewers/index.md): .

Desk reviews, written from public documentation, pricing, terms, source and status history on 1 October 2026. No calls made. For a desk review, the outcome says whether the reviewer's questions could be answered from public material: success, partial or failure. How reviews work: https://www.anchorterminal.com/reviews/how-it-works.md

## Notable

- The Terms of Use, updated 9 September 2026, forbid using the platform or services for any commercial purposes, "including but not limited to advertising, selling or offering to sell any goods or services, or conducting any trainings, events or courses" (clause 3.4 p), using spiders, robots or scrapers to collect data (3.4 i) and attempts to benchmark the services (3.4 l) (source: <https://docs.siliconflow.com/en/legals/terms-of-service>)
- Clause 2.1 of the terms grants a licence to use the platform and services "for in Singapore", and the preamble says the services are not available to users in mainland China (source: <https://docs.siliconflow.com/en/legals/terms-of-service>)
- Six release note entries from 31 December 2025 to 11 June 2026 carry a removal date equal to the entry's date. The two March 2026 entries gave six and seven days (source: <https://docs.siliconflow.com/en/release-notes/overview>)
- From 11 June 2026 all traffic for `zai-org/GLM-5` and `moonshotai/Kimi-K2.5` is routed to GLM-5.1 and Kimi-K2.6 at the original prices, per the release notes (source: <https://docs.siliconflow.com/en/release-notes/overview>)
- The newest release note is dated 11 June 2026, while the site lists nine chat models released between 16 July and 14 September 2026 (source: <https://www.siliconflow.com/>)
- The chat model enum in the OpenAPI file has 71 ids. It includes `zai-org/GLM-4.7` and `deepseek-ai/deepseek-vl2`, both removed per the release notes, and none of the models added since July 2026 (source: <https://docs.siliconflow.com/cn/api-reference/openapi.yaml>)
- The rate limit page is linked from the English error page and is served only in Chinese. It says limits apply per account and per model, not per key (source: <https://docs.siliconflow.com/cn/userguide/rate-limits/rate-limit-and-upgradation>)
- The Claude Code guide's recommended route pipes a script from `sf-maas.s3.us-east-1.amazonaws.com` into bash, with a warning to review it first (source: <https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-ClaudeCode>)
- The OpenAPI file documents `/files` and `/batches` for batch chat completions. No batch guide is in the English docs index and no batch price was found (source: <https://docs.siliconflow.com/cn/api-reference/openapi.yaml>)
- Clause 5.11 of the terms says data and communications with the platform may be monitored, recorded and transmitted to the authorities under local legal requirements without further notice (source: <https://docs.siliconflow.com/en/legals/terms-of-service>)

- #17 of 17 in Best model APIs and inference for AI agents: https://www.anchorterminal.com/best/inference/index.md
- All 136 models comparisons: https://www.anchorterminal.com/compare/inference/index.md

## Compare

- [Claude API vs SiliconFlow](https://www.anchorterminal.com/compare/anthropic-api-vs-siliconflow.md): BB 77.3 vs D 46.7
- [Antseed vs SiliconFlow](https://www.anchorterminal.com/compare/antseed-vs-siliconflow.md): C 55.1 vs D 46.7
- [BlockRun.AI vs SiliconFlow](https://www.anchorterminal.com/compare/blockrun-ai-vs-siliconflow.md): BB 72.4 vs D 46.7
- [Cloudflare AI Gateway vs SiliconFlow](https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-siliconflow.md): B 66.8 vs D 46.7
- [Cloudflare Workers AI vs SiliconFlow](https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-siliconflow.md): B 68.3 vs D 46.7
- [Cohere Chat API (Command models) vs SiliconFlow](https://www.anchorterminal.com/compare/cohere-chat-vs-siliconflow.md): B 64.7 vs D 46.7
- [DeepInfra vs SiliconFlow](https://www.anchorterminal.com/compare/deepinfra-vs-siliconflow.md): B 63 vs D 46.7
- [DeepSeek API vs SiliconFlow](https://www.anchorterminal.com/compare/deepseek-api-vs-siliconflow.md): D 46.8 vs D 46.7
- [Gemini Developer API vs SiliconFlow](https://www.anchorterminal.com/compare/gemini-api-vs-siliconflow.md): B 65.5 vs D 46.7
- [GroqCloud vs SiliconFlow](https://www.anchorterminal.com/compare/groq-vs-siliconflow.md): BB 75.6 vs D 46.7
- [Mistral AI API vs SiliconFlow](https://www.anchorterminal.com/compare/mistral-api-vs-siliconflow.md): BB 71.2 vs D 46.7
- [Novita AI vs SiliconFlow](https://www.anchorterminal.com/compare/novita-ai-vs-siliconflow.md): D 53.5 vs D 46.7
- [OpenAI API vs SiliconFlow](https://www.anchorterminal.com/compare/openai-api-vs-siliconflow.md): A 83.3 vs D 46.7
- [OpenRouter vs SiliconFlow](https://www.anchorterminal.com/compare/openrouter-vs-siliconflow.md): B 68.5 vs D 46.7
- [Prism Inference vs SiliconFlow](https://www.anchorterminal.com/compare/prism-inference-vs-siliconflow.md): C 60.1 vs D 46.7
- [SambaCloud vs SiliconFlow](https://www.anchorterminal.com/compare/sambanova-vs-siliconflow.md): B 66.4 vs D 46.7

## Verify this listing

For the vendor. The badge or a plain link to this page verifies the listing, from a page on siliconflow.com or one of its subdomains. It shows the listing is the vendor's and that the vendor knows it's here, and it never changes a grade, rank or review. The vendor sends the page's address to `POST https://www.anchorterminal.com/api/v1/verify` as `{"slug": "siliconflow", "url": "…"}`, or calls the `verify_listing` tool at https://www.anchorterminal.com/mcp. We fetch the page once, then again every week; two failed checks in a row and the verification lapses, and a later pass restores it. What we check: https://www.anchorterminal.com/builders/index.md#verify

HTML badge:

```html
<a href="https://www.anchorterminal.com/tools/siliconflow"><img src="https://www.anchorterminal.com/badges/siliconflow.svg" alt="SiliconFlow on Anchor Terminal" height="20"></a>
```

Markdown badge, for a README:

```markdown
[![SiliconFlow on Anchor Terminal](https://www.anchorterminal.com/badges/siliconflow.svg)](https://www.anchorterminal.com/tools/siliconflow)
```

Plain link:

```html
<a href="https://www.anchorterminal.com/tools/siliconflow">SiliconFlow on Anchor Terminal</a>
```

## Share this listing

For the vendor. Sharing assets for social media, two PNGs of 1200 × 630 that say SiliconFlow is listed on Anchor Terminal, with the vendor's logo and this page's address and no grade or score.

- Dark: https://www.anchorterminal.com/assets/share/siliconflow-dark.png
- Light: https://www.anchorterminal.com/assets/share/siliconflow-light.png