Qwen3.5-35B-A3B
Qwen3.5-35B-A3B is an efficient Mixture-of-Experts model from Alibaba's Qwen3.5 series with 35B total parameters and only 3B activated per token.
text + image + video → text · made by Alibaba
Sold by 20 ways
| Seller | Lane | Rate |
|---|---|---|
| Nous Research aggregator | standard | $0.08per Mtok in$0.75per Mtok out$0.12per Mtok cachedinference-api.nousresearch.com · read 2026-09-12 |
| OpenRouter 8 ways | $0.16per Mtok in$1.3per Mtok out$0.16per Mtok cached | |
| OpenRouter aggregator | darkbloom/fp4 | $0.08per Mtok in$0.75per Mtok outopenrouter.ai · read 2026-08-27 |
| OpenRouter aggregator | deepinfra/fp8 | $0.14per Mtok in$1per Mtok out$0.05per Mtok cachedopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregator | parasail/fp8 | $0.15per Mtok in$1per Mtok out$0.05per Mtok cachedopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregator | standard | $0.16per Mtok in$1.3per Mtok out$0.16per Mtok cachedopenrouter.ai · read 2026-09-15 |
| OpenRouter aggregator | alibaba | $0.16per Mtok in$1.3per Mtok outopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregator | atlas-cloud/fp8 | $0.23per Mtok in$1.8per Mtok out$0.23per Mtok cachedopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregator | siliconflow/fp8 | $0.24per Mtok in$1.8per Mtok out$0.15per Mtok cachedopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregatornot on the seller's list today | coreweave/fp8 | $0.25per Mtok in$1.25per Mtok out$0.25per Mtok cachedopenrouter.ai · read 2026-08-25 |
| OpenRouter aggregator | venice | $0.31per Mtok in$1.25per Mtok out$0.16per Mtok cachedopenrouter.ai · read 2026-08-25 |
| Requesty aggregatornot on the seller's list today | deepinfra flex | $0.1per Mtok in$0.72per Mtok out$0.036per Mtok cachedrouter.requesty.ai · read 2026-08-25 |
| Requesty aggregator | deepinfra | $0.14per Mtok in$1per Mtok out$0.05per Mtok cachedrouter.requesty.ai · read 2026-09-16 |
| DeepInfra aggregator | standard | $0.14per Mtok in$1per Mtok outapi.deepinfra.com · read 2026-08-25 |
| Hugging Face Inference Providers 3 ways | $0.25per Mtok in$2per Mtok out | |
| Hugging Face Inference Providers aggregator | deepinfra | $0.14per Mtok in$1per Mtok outrouter.huggingface.co · read 2026-08-25 |
| Hugging Face Inference Providers aggregator | standard | $0.25per Mtok in$2per Mtok outmodels.dev · read 2026-09-16 |
| Hugging Face Inference Providers aggregator | novita | $0.25per Mtok in$2per Mtok outrouter.huggingface.co · read 2026-08-25 |
| SiliconFlow aggregator | standard | $0.24per Mtok in$1.8per Mtok outmodels.dev · read 2026-09-16 |
| Novita AI aggregator | standard | $0.25per Mtok in$2per Mtok outapi.novita.ai · read 2026-08-25 |
| Venice AI aggregator | standard | $0.31per Mtok in$1.25per Mtok out$0.16per Mtok cachedmodels.dev · read 2026-09-16 |
| PPIO aggregator | standard | $1.6per Mtok in$12.8per Mtok outapi.ppinfra.com · read 2026-09-16 |
Measured 13 standings
| Place | Board | Metric | Score |
|---|---|---|---|
| 17thof 17 | SWE-rebench Leaderboard (default window 15/05/2026–01/07/2026, 111 problems) | Resolved Rate (%) | 17.1 |
| 23rdof 32 | AIME 2026 | Accuracy (± 95% CI) | 93.33 |
| 61stof 105 | Vectara Hallucination Leaderboard | Hallucination Rate | 10.5 |
| 66thof 123 | Lmca — Epoch AI | Score | 29.47 |
| 69thof 161 | Dtbench — Epoch AI | Accuracy | 80 |
| 96thof 313 | GPQA Diamond | mean_score | 83.46 |
| 122ndof 184 | Artificial Analysis · Coding Index | Index | 37 |
| 122ndof 177 | Critpt — Epoch AI | Accuracy (none) | 0.571 |
| 132ndof 222 | Chess puzzles — Epoch AI | mean_score (none) | 10 |
| 133rdof 290 | OTIS Mock AIME 2024-2025 | mean_score (none) | 70 |
| 142ndof 302 | Artificial Analysis · Intelligence Index | Artificial Analysis Intelligence Index | 15 |
| 150thof 168 | Scicode — Epoch AI | Score (none) | 29.282 |
| 378thof 656 | Epoch capabilities index — Epoch AI | ECI Score (none) | 142.54 |
About
Qwen3.5-35B-A3B — a text model from Alibaba, sold by 9 companies from $0.08 in and $0.75 out per million tokens, placed 96th of 313 on GPQA Diamond.
It takes text, images and video and returns text, with a context window of 262,144 tokens. It was published in February 2026. Its sellers say it can reason step by step and call a tool. The catalogue files it under reasoning and code. Nine companies sell it. The cheapest is $0.08 in and $0.75 out per million tokens at Nous Research. Beside the standard rate there are separately routed and regional lanes. It has been measured on 13 boards, and stands best at 96th of 313 on GPQA Diamond.
Every current figure
- Maker
- Alibaba
- Register
- model
- Takes
- text + image + video
- Returns
- text
- Context
- 262,144 tokens
- Published
- February 2026
- Parameters
- 36 billion
- Licence
- not read
- Sellers
- 9
- Maker's own price
- not read
- Price
- $0.08 in and $0.75 out per million tokens — Nous Research
- Boards
- 13
- Best place
- 96th of 313 — GPQA Diamond
Known as 3 names
Qwen/Qwen3.5-35B-A3Bdeepinfra/Qwen/Qwen3.5-35B-A3Bdeepinfra/Qwen/Qwen3.5-35B-A3B:flex