Data 2026-10-19 05:01 · USD/CNY 6.74
中文← Sinan Lab › Relay site comparisons › runapi.host vs tokenstation.top
Relay site comparisons
runapi.host vs tokenstation.top
As of 2026-10-19, runapi.host and tokenstation.top both sell 24 mainstream models: runapi.host has the lower effective price on 0, tokenstation.top on 24, and 0 are about the same. Across these shared models, runapi.host's median effective price is 60% of reference and tokenstation.top's is 5.1%.
Effective = the listed price converted into real dollars at the site's top-up rate and FX, per million output tokens. We make no recommendation; we just put both sites' measurements side by side.
Facts for both sites
| runapi.host | tokenstation.top | |
|---|---|---|
| Price profile | Near public price | Ultra-low |
| Median effective price of mainstream models (vs reference) | 60% (35 mainstream) | 5.1% (25 mainstream) |
| Models with quotes | 180 | 40 |
| Top-up rate | ¥6.75 per $1 | ¥0.091 per $1 |
| Group used for pricing | default ×1.00 | Claude code kiro ×2.80 and 3 groups in total (each model uses its lowest-multiplier public group) |
| 24-hour reachability | 98% | 100% |
| Median time to first byte | 2398 ms | 8719 ms |
| New-user sign-up | Open | Open |
| Panel | one-api | one-api |
| First indexed | 2026-09-02 | 2026-09-05 |
Mainstream models sold by both · 24
$ per million output tokens; small print is the listed price (site credits, unconverted).
| Model | runapi.host | tokenstation.top | Reference price | Lower effective price |
|---|---|---|---|---|
| Claude Opus 5.5 | 12.02listed $12.00 | 3.24listed $240.00 | 20.00 | tokenstation.top lower by 73% |
| Claude Fable 5.1 | 30.06listed $30.00 | 1.89listed $140.00 | 50.00 | tokenstation.top lower by 94% |
| GPT-6.1 Sol | 4.01listed $4.00 | 0.513listed $38.00 | 10.00 | tokenstation.top lower by 87% |
| GPT-6 Astra | 20.04listed $20.00 | 2.56listed $190.00 | 50.00 | tokenstation.top lower by 87% |
| GPT-6 Sol | 4.01listed $4.00 | 0.513listed $38.00 | 10.00 | tokenstation.top lower by 87% |
| Gemini 3.8 Flash | 1.58listed $1.57 | 0.192listed $14.25 | 4.12 | tokenstation.top lower by 88% |
| DeepSeek V4.1 Flash | 0.475listed $0.474 | 0.032listed $2.40 | 0.144 | tokenstation.top lower by 93% |
| DeepSeek V4 Flash | 1.07listed $1.07 | 0.032listed $2.40 | 0.182 | tokenstation.top lower by 97% |
| DeepSeek V4 Pro | 3.21listed $3.20 | 0.109listed $8.10 | 0.418 | tokenstation.top lower by 97% |
| GLM-5.3 | 3.33listed $3.32 | 0.227listed $16.80 | 1.00 | tokenstation.top lower by 93% |
| GLM-5.3 Flash | 0.333listed $0.332 | 0.023listed $1.68 | 0.200 | tokenstation.top lower by 93% |
| GLM-5.2 | 3.33listed $3.32 | 0.227listed $16.80 | 2.42 | tokenstation.top lower by 93% |
| Grok 4.7 | 1.80listed $1.80 | 0.227listed $16.80 | 6.00 | tokenstation.top lower by 87% |
| GPT-5.6 Sol | 12.02listed $12.00 | 1.54listed $114.00 | 20.00 | tokenstation.top lower by 87% |
| GPT-5.6 Terra | 4.81listed $4.80 | 0.615listed $45.60 | 12.00 | tokenstation.top lower by 87% |
| Claude Opus 4.8 | 15.03listed $15.00 | 0.945listed $70.00 | 25.00 | tokenstation.top lower by 94% |
| Claude Opus 4.6 | 15.03listed $15.00 | 0.945listed $70.00 | 25.00 | tokenstation.top lower by 94% |
| Claude Sonnet 4.6 | 9.02listed $9.00 | 0.567listed $42.00 | 15.00 | tokenstation.top lower by 94% |
| Claude Opus 4.7 | 15.03listed $15.00 | 0.945listed $70.00 | 25.00 | tokenstation.top lower by 94% |
| Claude Sonnet 5 | 6.01listed $6.00 | 0.567listed $42.00 | 10.00 | tokenstation.top lower by 91% |
| Claude Opus 5 | 15.03listed $15.00 | 0.945listed $70.00 | 25.00 | tokenstation.top lower by 94% |
| Claude Fable 5 | 30.06listed $30.00 | 1.89listed $140.00 | 50.00 | tokenstation.top lower by 94% |
| GPT-5.5 | 12.02listed $12.00 | 1.54listed $114.00 | 30.00 | tokenstation.top lower by 87% |
| Claude Haiku 4.5 | 3.01listed $3.00 | 0.189listed $14.00 | 5.00 | tokenstation.top lower by 94% |
This is an arithmetic ratio. It is not an accusation against the vendor, and a lower public source not yet indexed here may exist.
FAQ
- Which is cheaper, runapi.host or tokenstation.top?
- It depends on the model. Of the 24 mainstream models both sell, runapi.host is lower on 0, tokenstation.top is lower on 24 (e.g. Claude Opus 5.5、Claude Fable 5.1、GPT-6.1 Sol), and 0 are about the same. We only show effective prices and make no recommendation.
- How are the effective prices at runapi.host and tokenstation.top calculated?
- Effective = listed price × group multiplier × top-up rate ÷ FX (today 6.74). runapi.host: top-up rate ¥6.75 per $1, group default ×1.00; tokenstation.top: top-up rate ¥0.091 per $1, group Claude code kiro ×2.80 and 3 groups in total (each model uses its lowest-multiplier public group). A site's own page may show a member-group price by default.
- Which is easier to reach, runapi.host or tokenstation.top?
- Our probes over the last 24 hours: runapi.host reachable 98%, median time to first byte 2398 ms; tokenstation.top reachable 100%, median time to first byte 8719 ms. Reachability only says whether it can be reached, not model quality.
- Can you still sign up at both sites?
- runapi.host: Open; tokenstation.top: Open.
- How often is this data updated?
- Panel prices are fetched daily and reachability is probed hourly; every price can be traced to a fetch snapshot on each site's page.
More comparisonsrunapi.host vs runapi.corunapi.host vs aigcbest.toprunapi.host vs apidock.airunapi.host vs apilio.airunapi.host vs apiok.ccrunapi.host vs cherryin.netrunapi.host vs gptapi.asiarunapi.host vs magic666.toptokenstation.top vs runapi.cotokenstation.top vs aigcbest.toptokenstation.top vs apidock.aitokenstation.top vs apilio.aitokenstation.top vs apiok.cctokenstation.top vs cherryin.netAll comparisons →
One email a week
Subscribe to the Sinan weekly
This week's rankings, the price index, mainstream price changes, newly indexed and unreachable sites. Measurements only, no recommendations; unsubscribe any time.
Clear direction, grounded judgement.