Qwen3.8 2.4T A95B
AlibabaReleased Aug 12, 2026Open weights · Qwen3.8-Max Licenseqwen3.8-2.4t-a95b
- Reasoning
- yes, with effort levels
- Knowledge cutoff
- September 2025
- Context
- 262K tokens
- Max output
- 128k tokens
- Input
- text
- API features
- tools, structured outputs, JSON mode, seed, logprobs
- Alibaba's API
- yes, as qwen3.8-2.4t-a95b
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from Alibaba · cutoff: aiknowledgecutoff.com
The cheapest host, Novita AI, does not state its weights precision. Weights and context per host
Speed by provider
Output speed of Qwen3.8 2.4T A95B at each host, fastest first
- Together AI157 tok/s0.45 s to first token
- Venice136 tok/s0.75 s to first token
- SiliconFlow38 tok/s1.53 s to first token
- Alibaba34 tok/s2.01 s to first token
Speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · a snapshot of real traffic before the daily read, not a daily average · bars from zero.
Run it yourself
open weights, Qwen3.8-Max License · too large for the 6 workstations in our guideSelf-host guide ›- Parameters
- 2400B
- 95B active per token (MoE)
- Memory needed
- 1440 GB
- in INT4 · 2880 GB in INT8
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 1440 GB | multi-node, over 640 GB | smallest footprint, minor quality loss |
| INT8 (Q8) | 2880 GB | multi-node, over 640 GB | near-lossless |
| FP16 (full) | 5760 GB | multi-node, over 640 GB | reference quality |
At home: 1440 GB in INT4 is beyond any desktop or Mac in our machines table, so it means a multi-GPU server.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| DeepInfrafp4 (bf16) · 256K ctx (1M) | $2.00 | $6.00 | $0.20 | — | — | 100% | Oct 12 |
| Novita AIprecision not stated · 1M ctx · 128K out | $2.00 | $6.00 | $0.25 | — | — | — | Oct 23 |
| Alibabaprecision not read: host terms · 1M ctx · 128K out | $2.00 | $6.00 | $0.25 | 34 | 2.01 s | 100% | Oct 23 |
| SiliconFlowfp8 (bf16) · 1M ctx · 128K out | $2.00 | $6.00 | $0.25 | 38 | 1.53 s | 100% | Oct 23 |
| Modalprecision not stated · 1M ctx | $2.00 | $6.00 | $0.25 | — | — | 100% | Oct 13 |
| Together AIfp4 (bf16) · 1.01M ctx | $2.00 | $6.00 | $0.25 | 157 | 0.45 s | 100% | Oct 23 |
| Veniceprecision not read: host terms · 256K ctx (1M) · 64K out−20% | $2.00 | $6.00 | $0.25 | 136 | 0.75 s | 99.6% | Oct 23 |
7 of 7 providers · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · −N%: a discount off the host's list price · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| Novita AI | not stated by the host | 1M | 128K | yes | yes | yes |
| Alibaba | not read: host terms | 1M | 128K | yes | yes | yes |
| SiliconFlow | fp8 | 1M | 128K | yes | yes | yes |
| Modal | not stated by the host | 1M | not stated by the host | yes | yes | yes |
| DeepInfra | fp4 | 256K | not stated by the host | yes | yes | yes |
| Together AI | fp4 | 1.01M | not stated by the host | yes | yes | yes |
| Venice | not read: host terms | 256K | 64K | yes | no | yes |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
Qwen3.8 2.4T A95B spec sheet, field by field, with sources
Qwen3.8 2.4T A95B spec sheetmaker-declared
What Alibaba declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- 262,144 tokens
- Parameters
- 2,446,182,725,504≈ 2.45T
- Active experts
- 10 of 512 active per token
- Weight precision
- BF16
- License
- Qwen3.8-Max License — open weights, conditionalcurated · read 2026-09-07Scope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- not published by the maker≈ September 2025 third-party estimate — not the maker's figurecurated
- Modalities
- text
- Max output (official)
- not published by the makermaker's docs · read 2026-09-19
- Release date
- to verify≈ August 12, 2026 third-party catalogue — not the maker's figure
- Maker lifecycle status
- to verify
Also declared in the weights
- Architecture class
- Qwen3_5MoeForCausalLM
- Layers
- 92
- Vocabulary
- 248,320 tokens
- Weights first published
- August 8, 2026
- Weights last updated
- August 12, 2026
Qwen/Qwen3.8-2.4T-A95B · config.json @ 207bd68 · read 2026-10-08 · 6 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Sep 13 | Venice | $2.50 → $2.00 | $7.50 → $6.00 | −20% |
| Sep 5 | Together AI | $2.50 → $2.00 | $6.25 → $6.00 | −9% |
Every published change of input or output price, newest first · change on input + output · full series: pricing-history.csv.
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| Novita AI | $5.00 | $19.00 | $14.00 |
| Alibaba | $5.00 | $19.00 | $14.00 |
| SiliconFlow | $5.00 | $19.00 | $14.00 |
| Modal | $5.00 | $19.00 | $14.00 |
| DeepInfra | $5.00 | $19.00 | $14.00 |
| Together AI | $5.00 | $19.00 | $14.00 |
| Venice | $5.00 | $19.00 | $14.00 |
What each host bills for 1,000 requests of a representative job
Your own workload
Other Alibaba models · all its models, hosts and history ›
- Qwen3.8 27B16 providers · from $1.29
- Qwen3.6 35B A3B12 providers · from $1.00
- Qwen3 235B A22B Instruct 250710 providers · from $0.44
- Qwen3.5 397B A17B10 providers · from $3.45
- Qwen3.6 27B9 providers · from $2.25
- Qwen3.5-35B-A3B8 providers · from $1.14