Short Answer

Open-weight models are expected to maintain at least 72.7% of token volume on OpenRouter, consistent with current reported volumes of approximately 73% and no expected disruptions to this trend. The market prices this outcome at 90.0%.

1. Market Behavior & Drivers

This market's primary move was a direct repricing based on gateway data showing open-weight models already commanded a supermajority of inference token volume. The price gapped from 1.0% to 88.0% on October 27, reflecting reports that models like DeepSeek, Qwen, Llama, and Mistral accounted for approximately 73% of inference tokens on platforms like OpenRouter as of late October 2026. This data effectively resolved the market's core question near its inception.
The price has remained stable near its peak, supported by model-specific statistics. DeepSeek, a leading open-weight author, was reported to have a 26.5% share of text requests for the week beginning October 19. One of its models, DeepSeek V4.1 Flash 1017, held roughly 24.8% of total gateway tokens as of October 28. This concentration of volume in a single open-weight model author reinforced the market's high conviction.
  • Open-weight models consistently hold ~73% of token volume on OpenRouter.
  • No major model releases are expected, ensuring open-weight stability before November 1st.
  • DeepSeek V4.1 Flash's leading throughput supports high open-weight token share.

Who Wins and Why

Outcome Market Model Why
Above 70.7% 87.0% 92.3% Open-weight models are consistently reported to hold about 73% of token volume.
Above 69.2% 93.0% 94.7% Open-weight models are consistently reported to hold about 73% of token volume.
Above 71.3% 85.0% 92.3% Open-weight models are consistently reported to hold about 73% of token volume.
Above 73.9% 80.0% 86.1% Prediction markets indicate strong confidence for token share to exceed 73.9%.
Above 72.7% 90.0% 92.3% Open-weight models are consistently reported to hold about 73% of token volume.

Current Context

Open-weight models captured majority inference token volume on major gateways. These models, including DeepSeek, Qwen, Llama, and Mistral, accounted for approximately 73% of inference token volume on platforms like OpenRouter as of late October 2026 [1][2]. DeepSeek led model authors by token volume; its DeepSeek V4.1 Flash 1017 model held roughly 24.8% of total gateway tokens as of October 28, 2026 [1][3][4]. DeepSeek also topped text request counts with a 26.5% share during the week beginning October 19, 2026 [1][5]. Google followed with 20.1%, and OpenAI with 18.5% [1][5].
Open-weight models drive token growth, not revenue or advanced reasoning. While these models capture over 70% of token volume, they represent an estimated 23% of total spend as of October 2026 [6][7]. This disparity results from frontier proprietary models commanding significantly higher price points [6]. Industry-wide inference demand reached approximately 187 trillion tokens per week as of October 28, 2026 [2][8]. Open-weight models primarily fuel this growth through low-cost, high-throughput utility rather than through matching frontier reasoning capabilities [2][8].
Sources (8)
  1. 1LLM Rankingsopenrouter.ai
  2. 2Token Demand Index — AI Inference Consumption & Effective Price | Bargobargo.ai
  3. 3OpenRouter rankingstidelines.ai
  4. 4Model Migration, Normalized to Market Share — CodeSOTAort.fabryka.ai
  5. 5Chinese AI Models Lead OpenRouter Usage Rankings | LLMs Blogllms.blog
  6. 6Open weight takes 71.5% of tokens and 23% of spendvercel.com
  7. 7Enterprises Embrace Open Weights but Infrastructure Sellers Profit Mostfuturumgroup.com
  8. 8Open Weight Models Hit Half of AI Tokensfathom.news

2. Price Chart

Historical Price (Probability)

Outcome probability
Date

3. Significant Price Movements

Notable price changes detected in the chart, along with research into what caused each movement.

Outcome: Above 73.2%

📈 October 29, 2026: 84.0pp spike

Price increased from 1.0% to 85.0%

What happened: The provided research does not explicitly identify an 84.0 percentage point spike in the "Open source token share" market on October 29, 2026 [1]. Therefore, a primary driver for this specific price movement cannot be determined from the available sources. No social media activity, including posts from key figures or viral narratives, was found that would correlate with or explain such a movement [2][3][4][5][1]. Based on the absence of evidence for both the movement and related social media, social media was irrelevant in driving this purported market movement.

Outcome: Above 72.7%

📈 October 28, 2026: 88.0pp spike

Price increased from 0.0% to 88.0%

What happened: The reported 88.0 percentage point spike in the prediction market for "Open source token share Oct 26–Nov 1 (OpenRouter)" likely stemmed from a conflation of unrelated market data, not actual OpenRouter performance. The specific figures—an 88.0 percentage point difference and an "Above 72.7%" threshold—precisely match statistics from an October 2026 report on Shopify usage-based pricing apps, not OpenRouter's AI token share [6]. There is no evidence of such a spike or market share exceeding 72.7% for OpenRouter during that period; DeepSeek, the leading provider, held significantly lower market share [7][8][9][10]. Social media activity was irrelevant as a primary driver, as the movement appears to be rooted in misapplied data from traditional market analysis.

Outcome: Above 70.7%

📈 October 27, 2026: 83.0pp spike

Price increased from 5.0% to 88.0%

What happened: The provided web research indicates there is no evidence of an 83.0 percentage point (pp) spike in the "Open source token share Oct 26–Nov 1 (OpenRouter)" prediction market on October 27, 2026 [4]. The figure "83.0" likely references benchmark scores from other models, such as GPT-6 Astra's 83.0% on LiveBench 2.0, or misattributed prediction market volatility data, rather than an actual market share movement [4]. Consequently, without evidence of the reported price movement, social media cannot be identified as a primary driver or contributing factor for this non-existent event.
Sources (10)
  1. 12026-09-04:: AIGREGATORaigregator.news
  2. 2Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  3. 3Open source token share Oct 26–Nov 1 (OpenRouter) (Open-weight models market share - Oct 26, 2026): Odds, Chart & News | 4Search Predictions4search.com
  4. 4Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  5. 5https://headsupai.io/ai-news-and-updates/openrouterheadsupai.io
  6. 6Best Usage-based pricing Shopify Apps Analytics October 2026 | ShopDigestshopdigest.com
  7. 7LLM Rankingsopenrouter.ai
  8. 8AI Market Share 2026 - Provider Share of OpenRouter Usage | whatstrending.aiwhatstrending.ai
  9. 9Chinese AI Models Lead OpenRouter Usage Rankings | LLMs Blogllms.blog
  10. 10OpenRouter rankings, October 2026: which of GPT, DeepSeek, and Gemini is growing fastest, and how far apart are the API prices? - Vuncloudvuncloud.com

4. Market Data

Contract Snapshot

This market resolves to YES if the open source token share, as measured by OpenRouter between October 26 and November 1, 2026, is above 73.9%; otherwise, it resolves to NO. Trading for this market begins on November 2 at 10:00 AM EST, with the market expected to settle on or after November 1, 2026. The resolution relies solely on OpenRouter's data for the specified period, and no other special settlement conditions are detailed.

Available Contracts

Market options and current pricing

Outcome bucket Yes (price) No (price) Last trade probability
Above 69.2% $0.88 $0.83 93%
Above 72.7% $0.89 $0.61 90%
Above 70.7% $0.87 $0.77 87%
Above 71.3% $0.88 $0.95 85%
Above 71.8% $0.88 $0.89 85%
Above 73.2% $0.78 $0.64 85%
Above 72.3% $0.88 $0.95 80%
Above 73.9% $0.71 $0.72 80%
Above 75.3% $0.65 $1.00 3%

Market Discussion

Prediction markets, such as those on Kalshi and Public.com, tracked the "Open source token share" on OpenRouter for the week of October 26–November 1, 2026, with traders betting on whether open-weight models would exceed specific token share thresholds [1][2][3]. As of October 29, 2026, usage data indicated DeepSeek, an open-weight model, reportedly held a significant share of OpenRouter's token volume and text requests [4][5][6]. This market activity coincided with a broader trend in AI prediction markets in October 2026, shifting focus from "best model" metrics to "agent" capabilities [7].

Sources (7)
  1. 1Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  2. 2Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  3. 3Open source token share Oct 26–Nov 1 (OpenRouter) (Open-weight models market share - Oct 26, 2026): Odds, Chart & News | 4Search Predictions4search.com
  4. 4LLM Rankingsopenrouter.ai
  5. 5AI Market Share 2026 - Provider Share of OpenRouter Usage | whatstrending.aiwhatstrending.ai
  6. 6Model Migration, Normalized to Market Share — CodeSOTAort.fabryka.ai
  7. 7The 'Best Model' Race Is Over. Here's What Actually Matters. | ComputeLeapcomputeleap.com

5. Trust Index

Octagon Trust Index Kalshi 63 Caution

Primary risk· Trade quality

How it adds up
Integrity80% of score78Good

Market integrity is low (65), but Integrity averages all three scores, so the other two pull it up. Only a critically low score would cap the total.

Trade quality20% of score5Avoid
Trust score63Caution

Weighted blend with hard caps — a critically weak safety pillar, or a severe trading anomaly, caps the total regardless of the rest. Full methodology · About the Trust Index

Trust profile
Integrity4 screens run · 6 don't apply
Market integrity
65Caution

Outcome can be influenced

Thin-volume movesTrade signals · high confidence

“Above 75.3%” made a sharp jump with almost no trading behind it.

Screened clean
Large tradesNo single trade dominates the tape. Clear
Size patternsTape too thin Clear
2 screens not run
Outcome control · doesn't apply to this market type
Thin-leg conviction · doesn't apply to this market type
Info fairness
80Good

Structural prior — no screens apply

No suspicious activity on any screen

Screened clean
Information exposureNo concentrated group can know the outcome materially early. Clear
4 screens not run
Move vs. catalyst · doesn't apply to this market type
Longshot repricing · doesn't apply to this market type
Pre-broadcast certainty · doesn't apply to this market type
Cross-venue timing · doesn't apply to this market type
Resolution quality
92Strong

Resolution mechanism is dependable

No suspicious activity on any screen

Resolution rules read: Clear

Liquidity
5Very thin

Hard to fill near the screen price

8 contracts · worst first · 1 below the volume floor — not scored

Above 72.7%5
Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Above 70.7%5
Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Above 73.2%5
Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Above 73.9%5
Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Above 75.3%26
Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Above 69.2%—
No recent trading to measure

No recent trading to measure

Above 71.3%—
No recent trading to measure

No recent trading to measure

Above 71.8%—
No recent trading to measure

No recent trading to measure

Move quality
74Confirmed

Moves came with real volume

8 contracts · worst first · 1 below the volume floor — not scored

Above 75.3%34
Unconfirmed · medium confidence

Unconfirmed move — light volume and/or it has partly reverted.

Above 70.7%69
Mostly confirmed · high confidence

Volume is light/neutral for the move over 6h

Above 73.2%71
Confirmed · high confidence

Volume is light/neutral for the move over 6h

Above 72.7%72
Confirmed · high confidence

Volume is light/neutral for the move over 6h

Above 73.9%75
Confirmed · high confidence

Volume is light/neutral for the move over 6h

Above 69.2%—
Only 0 candle closes

Only 0 candle closes

Above 71.3%—
Only 0 candle closes

Only 0 candle closes

Above 71.8%—
Only 0 candle closes

Only 0 candle closes

Rule clarity
100Clear

Counts toward Resolution quality, not here

8 contracts · worst first · 1 below the volume floor — not scored

Above 69.2%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 72.7%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 70.7%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 71.3%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 71.8%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 73.2%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 73.9%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 75.3%100
Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Trade quality
5Avoid

Thin market — entering or exiting at fair value will be difficult.

Outcome quality 5 (notional-weighted across 5 scored legs)

1 below the volume floor — not scored

Above 69.2%Primary–Market scoreLiquidity–Move quality–Rule clarity100
Why this score
Liquidity
No recent trading to measure

No recent trading to measure

Move quality
Only 0 candle closes

Only 0 candle closes

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 72.7%5Market scoreLiquidity5Move quality72Rule clarity100
Why 5
Liquidity
5 · Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Move quality
72 · Confirmed · high confidence

Volume is light/neutral for the move over 6h

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 70.7%5Market scoreLiquidity5Move quality69Rule clarity100
Why 5
Liquidity
5 · Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Move quality
69 · Mostly confirmed · high confidence

Volume is light/neutral for the move over 6h

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 71.3%–Market scoreLiquidity–Move quality–Rule clarity100
Why this score
Liquidity
No recent trading to measure

No recent trading to measure

Move quality
Only 0 candle closes

Only 0 candle closes

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 71.8%–Market scoreLiquidity–Move quality–Rule clarity100
Why this score
Liquidity
No recent trading to measure

No recent trading to measure

Move quality
Only 0 candle closes

Only 0 candle closes

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 73.2%5Market scoreLiquidity5Move quality71Rule clarity100
Why 5
Liquidity
5 · Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Move quality
71 · Confirmed · high confidence

Volume is light/neutral for the move over 6h

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 73.9%5Market scoreLiquidity5Move quality75Rule clarity100
Why 5
Liquidity
5 · Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Move quality
75 · Confirmed · high confidence

Volume is light/neutral for the move over 6h

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear
Above 75.3%26Market scoreLiquidity26Move quality34Rule clarity100
Why 26
Liquidity
26 · Very thin · low confidence

Light recent trading — size may not fill near the screen price.

Move quality
34 · Unconfirmed · medium confidence

Unconfirmed move — light volume and/or it has partly reverted.

Rule clarity
100 · Clear · high confidence

All checks pass

objective outcome · Clearofficial source · Cleartime boundary · Clear

underwriting_v1.4 · computed Oct 29, 2026

6. What impact could a new model release from OpenAI or Google before November 1st have on the token share of open-weight models on OpenRouter?

OpenRouter Token Share ForecastExceed 69.2% (93% probability for Oct 26–Nov 1, 2026) [1][2][3]
Current OpenRouter Market Share (Open-weight)Approximately 72% (late October 2026) [4][5]
Latest Google Flash Series UpdateGemini 3.7 Flash (August 2026) [6][7][8]
No major model releases are expected, ensuring open-weight stability. In the final week of October 2026, major new model releases from OpenAI or Google are not anticipated, which suggests no significant disruption to the established dominance of open-weight models on OpenRouter is expected in the short term. OpenAI's most recent major models were GPT-5.6 and GPT-Rosalind earlier in 2026, while Google's latest update in its Flash series was Gemini 3.7 Flash in August 2026 [6][7][8].
Prediction markets confidently forecast continued open-weight model dominance on OpenRouter. High confidence is evident in prediction markets, which assign a 93% probability that open-weight models will maintain a token share exceeding 69.2% on OpenRouter for the period of October 26 to November 1, 2026 [1][2][3]. This share is projected to remain above 72%–73% [1][2][3]. This forecast aligns with the strong upward trend in token usage by open-weight models on OpenRouter, which recently reached approximately 72% market share as of late October 2026. Their prevalence is largely attributed to their cost-effectiveness for high-volume, single-step tasks [4][5].
Sources (8)
  1. 1Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  2. 2Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  3. 3Open source token share Oct 26–Nov 1 (OpenRouter) (Open-weight models market share - Oct 26, 2026): Odds, Chart & News | 4Search Predictions4search.com
  4. 4Open-Weight Models Dominate Token Usage, But Closed-...linkedin.com
  5. 5LLM Rankingsopenrouter.ai
  6. 6OpenAI to sell its top models on Google Cloud in latest...reuters.com
  7. 7የGPT‑5.6 ገንቢ መመሪያ | OpenAIopenai.com
  8. 8Gemini 3.7 Flash: our most intelligent workhorse modelblog.google

7. How does the distribution of token volume versus developer spend on OpenRouter confirm the utility-driven adoption of open-weight models like DeepSeek and Mistral?

Open-weight Token Volume ShareApproximately 72% (late 2026) [1]
Open-weight Developer Spend ShareApproximately 27% (late 2026) [1]
Predicted Open-source Token ShareExceeding 73% (Oct 26–Nov 1) [2][3][4]
Open-weight models capture significant token volume with lower developer spend. On OpenRouter, models such as DeepSeek and Mistral exemplify utility-driven adoption, processing approximately 72% of the total token volume by late 2026. Despite this high volume, these models account for a significantly lower roughly 27% of developer spend [1]. This divergence suggests that open-weight models serve as near-substitutes for high-volume, cost-sensitive, single-step tasks [5][1].
This dynamic creates a dual-structure market for AI models. The high token volume coupled with lower spend indicates that open-weight models are favored for their efficiency in numerous, less complex operations. Conversely, proprietary models maintain their pricing power for mission-critical, complex, or multi-step chain-of-thought workloads [5][1]. This fosters an ecosystem where cost-efficient open models capture the volume-heavy market segment, while proprietary providers secure higher-value, reliability-focused enterprise spend, confirming that open-weight models do not negate the business model of proprietary providers [1].
Prediction markets further confirm the sustained relevance of open-weight models. Forecasts regarding the 'Open source token share Oct 26–Nov 1 (OpenRouter)' expected the share to remain high, with probabilities favoring values exceeding 73% for that specific period [2][3][4].
Sources (5)
  1. 1Open-Weight Models Dominate Token Usage, But Closed-...linkedin.com
  2. 2Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  3. 3Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  4. 4Open source token share Oct 26–Nov 1 (OpenRouter) | PredictPediapredictpedia.com
  5. 5State of AI 2025: 100T Token LLM Usage Studyopenrouter.ai

8. How does DeepSeek's leading V4.1 Flash model compare to top proprietary models from Google and OpenAI in terms of token cost and request share on OpenRouter?

DeepSeek V4.1 Flash Input Token Cost$0.30/M input tokens [1][2][3]
DeepSeek V4.1 Flash Output Token Cost$1.20/M output tokens [1][2][3]
DeepSeek Predicted OpenRouter ShareExceed 27% (week of Oct 26, 2026) [4][5]
DeepSeek V4.1 Flash offers specific token pricing details for its checkpoint 1017 model. Pricing is set at $0.30 per million input tokens and $1.20 per million output tokens, with a significantly reduced rate of $0.006 per million tokens for cached input [1][2][3]. The provided facts do not contain information regarding the token cost for proprietary models from Google or OpenAI.
DeepSeek is predicted to lead OpenAI in OpenRouter market share for the week of October 26, 2026. DeepSeek's predicted request share on OpenRouter is expected to exceed 27% [4][5]. For the same period, OpenAI's market share is predicted to fall within the range of 16% to 19% [6][7]. Information on the request share for proprietary models from Google is not available in the provided facts.
Open-weight models are projected to dominate OpenRouter's token share by late October 2026. Prediction markets suggest a high probability, ranging from approximately 85% to 94%, that the overall open-weight model token share on OpenRouter will exceed 71% to 74% for the period of October 26 to November 1, 2026 [8][9][10].
Sources (10)
  1. 1DeepSeek V4.1 Flash - API Pricing & Benchmarksopenrouter.ai
  2. 2DeepSeek API Pricing (October 2026)benchlm.ai
  3. 3DeepSeek Review (2026): Pricing, Pros and Conssubconscious.dev
  4. 4DeepSeek market share this week Odds & Predictions - Public.compublic.com
  5. 5DeepSeek market share this week Prediction Market Results - Public.compublic.com
  6. 6OpenAI market share this week Odds & Predictions - Public.compublic.com
  7. 7OpenAI market share this week Prediction Market Results - Public.compublic.com
  8. 8Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  9. 9Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  10. 10Market Share Odds & Predictions 2026kalshi.com

9. What is the update frequency and data granularity of OpenRouter's rankings and the Bargo.ai Token Demand Index for tracking model market share in late October 2026?

OpenRouter Token Usage PeriodTrailing 7-day [1][2]
Bargo.ai Index Update FrequencyDaily [2][3]
Prediction Market Resolution Time10:00 AM ET on the following Monday [4][5]
OpenRouter's model rankings use continuously updated trailing 7-day token usage. These rankings display token usage figures based on a trailing seven-day period, which are continuously updated and become final for a given "week" once the subsequent week's data is initially displayed [1][2]. Although the raw data is granular at the user-model-day level, OpenRouter's public ranking dashboards consolidate this information into weekly totals [6]. An OpenRouter "week" is precisely defined as spanning from Monday 12:00 AM UTC to Sunday 11:59 PM UTC [1][2].
Bargo.ai's index updates daily, utilizing OpenRouter's ranking data. The Bargo.ai Token Demand Index provides daily updates and is constructed using OpenRouter's public model rankings, incorporating both trailing-7-day token volume and pricing data [2][3]. This index categorizes models into various provider groups, including Claude, OpenAI, Google, xAI, Open-source, and other unspecified categories [2][3]. Furthermore, a prediction market focused on "Open source token share Oct 26–Nov 1 (OpenRouter)" will resolve based on the percentage of tokens processed by open-weight models for the week of October 26, 2026. The official value for this market will be the figure reported at 10:00 AM ET on Monday, November 2, 2026 [4][5].
Sources (6)
  1. 1OpenRouter AI token usage Oct 26 to Nov 22? - Polymarket Quoten | CoinRithmcoinrithm.com
  2. 2Token Demand Index — AI Inference Consumption & Effective Price | Bargobargo.ai
  3. 3https://www.bargo.ai/llms.txtbargo.ai
  4. 4Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  5. 5Open source token share Oct 26–Nov 1 (OpenRouter) | PredictPediapredictpedia.com
  6. 6AI Premiumarxiv.org

10. What usage data from OpenRouter and Bargo.ai validates the claim that Chinese models like DeepSeek and Qwen are leading total token throughput?

DeepSeek V4.1 Flash Tokens56.2T tokens (OpenRouter, trailing 7 days) [1]
Open-source Models Token Volume Share73.1% (all tracked inference tokens on OpenRouter) [2]
DeepSeek OpenRouter Request Share26.5% (week beginning Oct 19, 2026) [1]
DeepSeek's V4.1 Flash demonstrates leading token throughput on OpenRouter. DeepSeek V4.1 Flash variants showed substantial usage on OpenRouter, processing 56.2 trillion tokens in the trailing seven days as of October 28, 2026, which validates DeepSeek's leading position in total token throughput [1]. This volume significantly surpasses competitors, highlighting its dominance in the inference market [1].
Open-source models collectively command a substantial share of inference volume. Open-source and open-weight models together hold 73.1% of all tracked inference token volume on OpenRouter, according to Bargo.ai's Token Demand Index as of October 28, 2026 [2]. Within this segment, DeepSeek maintained a leading position with a 26.5% share of OpenRouter requests for the week beginning October 19, 2026, while Qwen accounted for 4.7% of requests during the same period [1]. Prediction markets, such as "Open source token share Oct 26–Nov 1 (OpenRouter)," further indicate strong market sentiment that open-weight models will maintain a share above 69.2%, aligning with the observed usage data [3][4][5]. While DeepSeek's leading token throughput is supported, the research does not provide specific data to validate Qwen's individual leadership in total token throughput [1][2].
Sources (5)
  1. 1LLM Rankings | OpenRouteropenrouter.ai
  2. 2Token Demand Index — AI Inference Consumption & Effective Price | Bargobargo.ai
  3. 3Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions 2026kalshi.com
  4. 4Open source token share Oct 26–Nov 1 (OpenRouter) Odds & Predictions - Public.compublic.com
  5. 5Open source token share Oct 26–Nov 1 (OpenRouter) (Open-weight models market share - Oct 26, 2026): Odds, Chart & News | 4Search Predictions4search.com

11. What Could Change the Odds

Key Catalysts

Open-weight models processed over 70% of tokens on major AI gateways, such as OpenRouter, as of October 2026 [1][2][3]. Despite this, open-weight models capture significantly less than 30% of total AI spending, indicating proprietary frontier models retain high-value, complex workloads [1][2][3]. The shift toward open-weight models is driven by high-volume, routine, and cost-sensitive enterprise and developer tasks [2][3][4]. Prediction markets for the week of October 26–November 1, 2026, indicate strong confidence, over 85-94% probability, in open-weight models maintaining a high token share above 72.7-73.9% [1][5][6].
Frontier AI labs, including OpenAI, Anthropic, and Google, maintain pricing power through superior dependability and reasoning capabilities for multi-step tasks [2][3][4]. Chinese AI models, including DeepSeek, GLM, and Hunyuan, have significantly increased their market share in 2026, contributing to the dominance of open-weight usage [2][7][3][8]. However, enterprise users are increasingly sensitive to country-of-origin and export-control exposure, which may influence future market shifts [2][7][3][8]. Bullish catalysts for frontier AI labs include enterprise growth and agentic reliability [3]. Bearish catalysts for the token-share dominance of open-weight models include potential enterprise shifts toward sovereign or more "dependable" architectures and regulatory and compliance pressures [3].

Key Dates & Catalysts

  • Expiration: November 02, 2026
  • Closes: November 02, 2026
Sources (8)
  1. 1Open source token share Oct 26–Nov 1 (OpenRouter) - AIkalshi.com
  2. 2Open weight takes 71.5% of tokens and 23% of spendvercel.com
  3. 3Enterprises Embrace Open Weights but Infrastructure Sellers Profit Mostfuturumgroup.com
  4. 420VC: Will OpenRouter Sell for $10BN to Stripe? | Why Chinese Open Models Are Beating America—and What Happens Next | Why Enterprises Are More Fearful of Anthropic and OpenAI Than China | Is the Routing Layer Becoming a Commodity with Alex Atallah - The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitchgetpodcast.com
  5. 5Open source token share Oct 19–25 (OpenRouter) Prediction Market Results - Public.compublic.com
  6. 6OpenAI market share this week Odds & Predictions 2026kalshi.com
  7. 7Enterprises Embrace Open Weights but Infrastructure Sellers Profit Most_AIONP202610futurumgroup.com
  8. 8The State of Open Source AI — v1.1 · September 2026stateofopensource.ai

13. Historical Resolutions

Historical Resolutions: 9 markets in this series

Outcomes: 5 resolved YES, 4 resolved NO

Recent resolutions:

  • KXOPENSOURCESHAREOR-26OCT26-76: NO (Oct 26, 2026)
  • KXOPENSOURCESHAREOR-26OCT26-75: NO (Oct 26, 2026)
  • KXOPENSOURCESHAREOR-26OCT26-74: NO (Oct 26, 2026)
  • KXOPENSOURCESHAREOR-26OCT26-73: NO (Oct 26, 2026)
  • KXOPENSOURCESHAREOR-26OCT26-72: YES (Oct 26, 2026)