Qwen3 235B A22B Instruct 2507
Sold by 5 sellers, timed at 33 tok/s, on 6 Design Arena boards, and not scored by Artificial Analysis.
retrieved sha256 · 0bb53a77713f
The same weights served by 4 providers, each at its own list price. OpenRouter's rate is its routed rate across whichever of them it uses, and it need not match any one of them.
What sellers charge
5 sellers · 2.08x from cheapest to dearestOne tick per seller, on a log scale of input price per million from $0.072 at DeepInfra to $0.15 at Venice. 3 sellers undercut the $0.14 listed; 1 is dearer. A faint tick declares no number format. Every seller is named on the endpoint ledger below.
Measured speed
20,395 requests · 5 endpoints · 30-minute windowOpenRouter'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
5 priced · 5 timed · cheapest first- DeepInfradeepinfra/fp8/flex$0.072$0.44not listedfp8236knot matchednonenonenot matched100%
- DeepInfradeepinfra/fp8$0.09$0.55not listedfp816knot matchednonenonenot matched92.89%
- DeepInfradeepinfra/fp8/priority$0.135$0.825not listedfp8236knot matchednonenonenot matched97.41%
- Parasailparasail/fp8 · listed$0.14$0.80$0.05fp8118k39 tok/s25597ms9,774 req100%
- Venicevenice/fp8$0.15$0.75not listedfp816k29 tok/s13700ms2,876 req99.77%
- DeepInfranot matchednot matchednot matchednot matchedno offer28 tok/s14421ms6,385 reqno offer
- DeepInfranot matchednot matchednot matchednot matchedno offer12 tok/s51.42s16 reqno offer
- DeepInfranot matchednot matchednot matchednot matchedno offer22 tok/s11600ms1,344 reqno offer
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: 4 of the 5 timed routes are slower than it, and the middle route runs at 28 tok/s. The widest median-to-p90 gap is DeepInfra's, 2.4x.
DeepInfra publishes 3 offers and had 3 routes timed. The two documents share no route id, so those offers and routes are printed apart rather than paired.
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 . 7 of the 7 sellers with a reading fell under 99% on at least one of the 23 days held, 23 Oct 2026 to 21 Nov 2026: GMICloud at 0.00% on 23 Oct 2026, DeepInfra (deepinfra/fp8/flex) at 40.00% on 19 Nov 2026, DeepInfra (deepinfra/fp8) at 79.65% on 17 Nov 2026, Nebius at 85.04% on 12 Nov 2026, and 3 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
17 parameters · 5 sellers- frequency_penaltyevery seller
- max_tokensevery seller
- presence_penaltyevery seller
- response_formatevery seller
- stopevery seller
- structured_outputsevery seller
- temperatureevery seller
- tool_choiceevery seller
- toolsevery seller
- top_kevery seller
- top_pevery seller
- logit_bias4 of 5not Venice
- repetition_penalty4 of 5not Venice
- seed4 of 5not Venice
- min_p3 of 5not Parasail and Venice
- logprobs1 of 5only Parasail
- top_logprobs1 of 5only Parasail
11 of these 17 are accepted by every one of the 5 sellers that publish a list; the other 6 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 modelsBlended $0.305 per million tokens, 3 in to 1 out: cheaper than 244 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 boardEach 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.
- 3D scenes · Models974 Elowin rate 41.1% · source rank 123the board, 113 ranked
- Code · Models1027 Elowin rate 42.6% · source rank 124the board, 120 ranked
- Data visualisation · Models1066 Elowin rate 49.0% · source rank 111the board, 110 ranked
- Game development · Models940 Elowin rate 35.0% · source rank 137the board, 120 ranked
- UI components · Models954 Elowin rate 38.4% · source rank 127the board, 116 ranked
- Websites · Models1049 Elowin rate 43.6% · source rank 126the board, 126 ranked
Position 134 of the 258 models Epoch has placed, which is not the population of either board above, on a scale anchored at Claude 3.5 Sonnet 130 and GPT-5 150. Epoch states that absolute values are meaningless by themselves and that comparisons between models are what the index is for, that values are not linearly related to benchmark accuracy, and that a highly specialised model may score low while being capable in its own domain.
Benchmark results
none of 6 boardsNone of the 6 openly licensed boards Vitna mirrors has published a result for Qwen3 235B A22B Instruct 2507 under a name that matches its listing, and names are matched to the catalog exactly and never guessed. Every board, with its licence and what its publisher measures, is on /benchmarks.
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.
Source: the Epoch AI Capabilities Index (epoch.ai/eci), as of 20 Nov 2026. 258 models placed over 2917 benchmark results.
Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.
Vitna ran none of these evaluations and adjusted none of these figures.
