Get started

Qwen3.8 27B (batch)

Sold by 24 sellers, timed at 115 tok/s, scored 20.1 on the Intelligence Index, and with results on 1 openly licensed board.

retrieved sha256 · 0bb53a77713f

Models/Qwen/Qwen3.8 27B (batch)USD per 1M tokens · as OpenRouter routes it
qwen/qwen3.8-27b:batchQwen · batch variant · Qwen tokenizer
OpenRouter list price · the routed rate, as the catalog carried itretrieved 17 Nov 16:29 UTC from openrouter.ai ↗
Input$0.16per 1M tokens
Output$2.00per 1M tokens
Cached input$0.04per 1M tokens
Context262k262,144 tokens
What the list says · the fields OpenRouter publishes for this id12 fields
AuthorQwen
Listed
Canonical idqwen/qwen3.8-27b-20260814
Context262,144 tokens
Max output235,929 tokens
Modalitiestext, image, video → text
TokenizerQwen
Reasoningoptional, on by default · effort xhigh, medium, low, none, xhigh by default
Moderationnot moderated by OpenRouter
Variantbatch
Other listingbase listing
What the weights say · read from the model's own repository on Hugging Faceretrieved from huggingface.co ↗
Licenceapache-2.0
Parameters27.8B · 27,781,427,952 declared
Model classQwen3_5ForConditionalGeneration
Downloadnot gated
Published

What sellers charge

24 sellers · 28.29x from cheapest to dearest

One tick per seller, on a log scale of input price per million from $0.035 at Darkbloom to $0.99 at Cerebras. 11 sellers undercut the $0.16 listed; 13 are dearer. A faint tick declares no number format. Every seller is named on the endpoint ledger below.

The same sellers grouped by the number format each declares, on the same scale. The fp4 group alone spans 76% of the range, so the cheap end is not cheap for its number format. A declaration is checked by nobody: this groups the rows and certifies none.

Measured speed

313,601 requests · 24 endpoints · 30-minute window
Throughput115 tok/sMedian output rate, weighted across the 24 endpoints serving it by the requests each answered.the speed board
First token2.1sMedian wait before the first token, over the same window and weighted the same way.how Vitna reads a figure
Endpoint spread9 to 279The slowest and the fastest route serving this model, in tokens per second. Speed is a property of the route, not of the weights.the board, fastest first

OpenRouter's own timing of its routed traffic, weighted by requests, over one 30-minute window read . Every route is on the endpoint ledger below.

Every endpoint

24 priced · 24 timed · cheapest first
  1. Darkbloomdarkbloom/fp4$0.035$1.50$0.0175fp4131k46 tok/s223.17s1,127 req99.54%
  2. Chuteschutes/fp8$0.039$1.25$0.0039fp866k22 tok/s46.77s40,936 req100%
  3. DekaLLMdekallm$0.04$2.50$0.02not declared236k120 tok/s65500ms8,404 req99.86%
  4. Ionstreamionstream/fp8$0.04$1.95$0.02fp866k9 tok/s811.7s14,885 req100%
  5. Modalmodal$0.04$2.30$0.02not declared236k177 tok/s98328ms22,998 req100%
  6. Near AInear-ai/fp8$0.04$1.35$0.018fp866k44 tok/s111.9s41,679 req100%
  7. Waferwafer/fp8$0.06$1.96$0.059fp8236k150 tok/s119646ms141,433 req100%
  8. Mancer 2mancer/fp8$0.125$1.50$0.03fp8236k51 tok/s2431.4s48 req89.24%
  9. DeepInfradeepinfra/fp8 · 25% off$0.15$1.875$0.0375fp8236k52 tok/s37538ms6,293 req100%
  10. Rekareka/fp8$0.15$2.50$0.05fp8236k59 tok/s251.06s1,260 req100%
  11. Tenstorrenttenstorrent/fp8$0.15$1.60$0.05fp8236k122 tok/s72399ms5,126 req99.63%
  12. StreamLakestreamlake · 65% off$0.175$1.05$0.035not declared128k53 tok/s26820ms221 req100%
  13. ModelRunmodelrun/fp4$0.20$2.50$0.09fp4236k193 tok/s135260ms10,408 req100%
  14. AkashMLakashml/fp8$0.225$1.98$0.05fp8131k40 tok/s27855ms346 req99.91%
  15. Parasailparasail/fp8$0.24$2.20$0.05fp8236k93 tok/s61547ms1,607 req99.92%
  16. CoreWeavecoreweave/fp8$0.40$3.00$0.15fp8236k69 tok/s51274ms3,969 req100%
  17. Novitanovita$0.42$3.00$0.085not declared131k44 tok/s71.02s15 req100%
  18. Alibabaalibaba$0.425$2.55$0.085not declared131k80 tok/s39.41.48s209 req100%
  19. SiliconFlowsiliconflow/fp8$0.425$2.55$0.085fp8236k77 tok/s431.04s26 req100%
  20. Cloudflarecloudflare$0.45$3.20$0.05not declared236k29 tok/s82.78s40 req100%
  21. Nebiusnebius/fp4$0.45$3.00$0.10fp4900k279 tok/s86674ms10,271 req100%
  22. PrimeIntellectprimeintellect/fp8$0.45$3.20$0.05fp866k113 tok/s41309ms266 req100%
  23. Venicevenice/fp8$0.45$3.20not listedfp866k64 tok/s43783ms137 req100%
  24. Cerebrascerebras/fp16$0.99$1.49$0.99fp1633k270 tok/s63.1355ms1,897 req100%

Each track runs from a route's p90, the rate nine in ten of its requests met or beat, up to its median. Weighted by requests, the figure above reports where the traffic went rather than what one route does: 17 of the 24 timed routes are slower than it, and the middle route runs at 66.5 tok/s. The widest median-to-p90 gap is Novita's, 6.29x.

Uptime is OpenRouter's own figure for each seller over the trailing day, read once a day, and the column prints the latest reading, taken . 24 of the 29 sellers with a reading fell under 99% on at least one of the 23 days held, 23 Oct 2026 to 21 Nov 2026: GMICloud (gmicloud/bf16) at 55.17% on 10 Nov 2026, Darkbloom at 75.12% on 6 Nov 2026, Cloudflare at 76.47% on 18 Nov 2026, GMICloud (gmicloud/fp8) at 76.78% on 14 Nov 2026, and 20 more. Source: OpenRouter's public endpoints API, one reading a day for each seller: 1,211 sellers carried one on the latest of the 23 days held.

Source: OpenRouter's public endpoints API. 1257 sellers across 359 models when the mirror ran, 186 of them served by more than one. List prices, not what anybody paid; a discount is the seller's own. Read . Speed is OpenRouter's own timing over one 30-minute window, read .

What each seller accepts

21 parameters · 24 sellers
  1. include_reasoningevery seller
  2. max_tokensevery seller
  3. reasoningevery seller
  4. reasoning_effortevery seller
  5. temperatureevery seller
  6. tool_choiceevery seller
  7. toolsevery seller
  8. top_pevery seller
  9. response_format23 of 24not Cerebras
  10. stop23 of 24not SiliconFlow
  11. frequency_penalty20 of 24
  12. presence_penalty20 of 24
  13. top_k20 of 24
  14. seed19 of 24
  15. structured_outputs19 of 24
  16. logprobs17 of 24
  17. top_logprobs17 of 24
  18. repetition_penalty14 of 24
  19. logit_bias11 of 24
  20. min_p6 of 24
  21. top_a2 of 24only Mancer 2 and PrimeIntellect

8 of these 21 are accepted by every one of the 24 sellers that publish a list; the other 13 are declared by some sellers and not others, so what a request sending one of them gets depends on which seller serves it. Each list is the seller's own declaration in OpenRouter's endpoints API, read , and nothing here tested one.

Price position

342 priced models

Blended $0.62 per million tokens, 3 in to 1 out: cheaper than 194 of the 342 priced models in the catalog. Each bar is one stretch of the catalog on a logarithmic price scale; the mark is this listing.

Measured quality

145 models with an index · 132 on a design arena board

Each mark is one board this model stands on, placed at the share of that board it scores above, and led to its number in the list under the rule. The columns and the ledger below carry the figures.

Intelligence Index20.1Position 39 of 58, above 33% of the models scored.the intelligence leaderboard
Coding Index68.1Position 31 of 118, above 74% of the models scored.the coding leaderboard
Agentic Index45.8Position 13 of 94, above 86% of the models scored.the agentic leaderboard

Not ranked in any Design Arena category.

Epoch Capabilities Index · a fit by a third party over every benchmark this model has results onthe whole index, with intervals

Epoch AI has not placed Qwen3.8 27B (batch) on its Capabilities Index, or places it under a name the pairing does not reach. Epoch places a model once four of its benchmarks carry a result.

Benchmark results

1 of 6 boards · 23 results
LiveBench · publisher-run, Apache-2.023 of 23 tasks · 71 runs scored on each task · effort unstated

Percent correct on the task the row names, higher is better. All 23 published columns; there is no average to carry, because the site computes its headline in the browser from a weighting this does not reimplement. LiveBench names this run qwen3.8-27b, with no effort setting written on it. Read from livebench.ai ↗, which dates its own data . LiveBench, Apache-2.0. Results produced by the LiveBench maintainers on questions refreshed monthly.

1 of the 6 openly licensed boards Vitna mirrors carries a result for Qwen3.8 27B (batch), each printed as its publisher published it and none added to another. Epoch AI Benchmarking Hub, MTEB, HELM, Vectara hallucination leaderboard and Aider polyglot have not published one under a name that matches its listing, and names are matched to the catalog exactly and never guessed. A task that is a link opens its column on /capability, with every run scored on it; an effort that is a link opens its run's row there, where the run is ranked on Vitna's scale. Every board, with its licence and what its publisher measures, is on /benchmarks.

Sources

Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Source: each model's own Hugging Face repository (huggingface.co), as of 20 Nov 2026. 162 repositories read, 4 did not answer.
Licence, parameter count, model class and gating: each model's own Hugging Face repository, the one OpenRouter names for the listing.
A licence is the tag the uploader put on the repository, read from nothing else and verified by nobody.
A parameter count is the total the repository's safetensors files declare; for a mixture of experts that is every expert, not the share one token runs through.
Vitna ran none of these evaluations and adjusted none of these figures.