SmophyAI

Published daily · Next update 02:00 UTC

AI Value Over Time · tracked daily

See what AI actually costs - over time

Every AI benchmark site shows you a snapshot. We track quality-per-dollar and real usage share every day - so you can see whether a model is getting better, cheaper, or more popular, not just where it stands today. No invented index: every number here is a real, sourced quantity or a plain ratio.

For a comparison of API prices and cost per solved task, read our guide to the best AI model for the money in October 2026.

Models tracked
404
Best value (top 10 usage)
DeepSeek: DeepSeek V4 Flash 0423
History
125 days
Latest data
November 6, 2026

Quality and usage aren't the whole picture - these track what running a model actually feels like in production, updated every day.

Uptime & latency

Provider reliability

  1. 1. DeepSeek: DeepSeek V4 Flash 0423100.0%
  2. 2. DeepSeek: DeepSeek V4.1 Flash (batch)100.0%
  3. 3. Tencent: Hy3100.0%

DeepSeek: DeepSeek V4 Flash 0423 tops the list at 100.0% best-provider uptime.

Full reliability tracker →

Real measured throughput

Fastest models

  1. 1. Qwen: Qwen3.8 27B (batch)835 tok/s
  2. 2. OpenAI: gpt-oss-120b (batch)813 tok/s
  3. 3. Z.ai: GLM 5.2528 tok/s

Qwen: Qwen3.8 27B (batch) leads at 835 tokens/second via Cerebras.

Full throughput ranking →

Daily rank changes

Latest leaderboard movement

  1. 1. Xiaomi: MiMo-V2.6-Pro↑ 5

November 6, 2026: DeepSeek: DeepSeek V4.1 Flash (batch) led real usage share. Xiaomi: MiMo-V2.6-Pro moved 5 places up.

Full daily changelog →

Quality per dollar, over time

Top 8 by usage · tracked daily since August 9, 2026

As of October 28, 2026, DeepSeek: DeepSeek V4 Flash 0423 delivers the best measured quality per dollar among widely-used models, at 230.59 intelligence-index points per dollar ($0.15/1M tokens blended). This chart is tracked daily - first-party history begins August 9, 2026 and grows by one point every night; it is not a static snapshot.

  1. DeepSeek: DeepSeek V4 Flash 0423 - 230.59 pts/$ · $0.15/1M
  2. Xiaomi: MiMo-V2.5 - 124.57 pts/$ · $0.18/1M
  3. Z.ai: GLM 5.3 Flash (batch) - 117.05 pts/$ · $0.24/1M
  4. DeepSeek: DeepSeek V4.1 Flash (batch) - 108.95 pts/$ · $0.26/1M
  5. OpenAI: GPT-5.6 Luna (batch) - 52.89 pts/$ · $0.45/1M
  6. Z.ai: GLM 5.2 - 19.01 pts/$ · $1.77/1M
  7. DeepSeek: DeepSeek V4 Pro 0423 - 18.44 pts/$ · $1.25/1M
  8. NVIDIA: Nemotron 3 Ultra (free) - 11.24 pts/$ · $1.05/1M
1118.9839.2559.4279.70.0Intelligence index per dollarAug 9Sep 22Nov 7
  • DeepSeek: DeepSeek V4 Flash 0423
  • DeepSeek: DeepSeek V4 Pro 0423
  • DeepSeek: DeepSeek V4.1 Flash (batch)
  • NVIDIA: Nemotron 3 Ultra (free)
  • OpenAI: GPT-5.6 Luna (batch)
  • Xiaomi: MiMo-V2.5
  • Z.ai: GLM 5.2
  • Z.ai: GLM 5.3 Flash (batch)

Index date November 6, 2026Data through 00:00 UTC November 7, 2026Calculated 02:00 UTC

How this is calculated

Quality per dollar= Artificial Analysis Intelligence Index ÷ blended price per 1M tokens (75% prompt / 25% completion). No weighting, no composite - a plain ratio of two real, independently sourced numbers, computed fresh from that day's pricing and benchmark snapshot. The Intelligence Index already measures per-response quality (accuracy across a benchmark suite of reasoning, coding, and knowledge tasks) - it is not response volume, so this ratio isn't "cheap model, more tries." A small quality gap next to a huge price gap is what drives a cheap model to the top.

What this ratio can't tell you:

  • Consistency - whether a model reliably hits the same quality run to run, not just on average.
  • Cost of a wrong answer in your specific use case - a bad response in production code is not the same as a bad response in casual chat.
  • Whether a task needs multiple attempts to reach a usable result.

This is a real, industry-standard way to pre-filter candidates by cost-efficiency - not a final verdict. For high-stakes tasks, weigh the absolute Intelligence Index (shown alongside every ratio below) more heavily than the ratio itself, and check a model's measured endpoint uptime for its consistency track record.

Top 15 models by quality per dollar

Intelligence Index points per dollar of blended price - the 15 best-value models among all 96 with a published quality score. This is a ratio, not a capability ranking: an ultra-cheap model with modest absolute quality can outscore a much stronger, pricier one, purely because price sits near zero. Each bar shows its own absolute Intelligence Index and price alongside the ratio, so you can see which is driving the number.

01OpenAI: gpt-oss-120b (batch)692.5pts/$Quality 11.6 · $0.02/1M
02inclusionAI: Ling 3.1 Flash608.9pts/$Quality 41.1 · $0.07/1M
03Upstage: Solar Mini 4275.4pts/$Quality 24.1 · $0.09/1M
04Anthropic: Claude Haiku 5.5 (batch)182.5pts/$Quality 36.5 · $0.20/1M
05Inception: Mercury 2.5182.2pts/$Quality 12.3 · $0.07/1M
06OpenAI: gpt-oss-20b (batch)163.6pts/$Quality 9.0 · $0.05/1M
07Xiaomi: MiMo-V2.6-Flash152.6pts/$Quality 26.7 · $0.18/1M
08OpenAI: GPT-6 Luna (batch)139.5pts/$Quality 27.9 · $0.20/1M
09Z.ai: GLM 5.3 Flash (batch)117.1pts/$Quality 27.8 · $0.24/1M
10MiniMax: MiniMax M3.1 Flash110.9pts/$Quality 29.1 · $0.26/1M
11DeepSeek: DeepSeek V4.1 Flash (batch)109.0pts/$Quality 28.6 · $0.26/1M
12Upstage: Solar Pro 4107.4pts/$Quality 28.2 · $0.26/1M
13NVIDIA: Nemotron 3 Nano 30B A3B84.8pts/$Quality 8.9 · $0.10/1M
14NVIDIA: Nemotron 3 Super81.9pts/$Quality 12.8 · $0.16/1M
15Z.ai: GLM 5.3 (batch)78.8pts/$Quality 31.7 · $0.40/1M
Download chartFull value ranking →

Open vs. closed AI, tracked daily

125 days of real usage history

Open-weight models now handle 78% of real token volume on OpenRouter - up 8.9 points since July 5, 2026. No other site tracks this as a living, dated trend.

100%75%50%25%0%Share of usageJul 5Sep 5Nov 6
  • Open-source
  • Closed
Full open vs. closed breakdown, with the quality gap →

Best AI model for…

Ranked by real usage share for that specific task, not overall popularity.

code

Code Generation

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)39%
  2. 2. Z.ai: GLM 5.3 Flash (batch)6%
  3. 3. Crouton Alpha5%

DeepSeek: DeepSeek V4.1 Flash (batch) leads with 39% of task usage, 33 points ahead of Z.ai: GLM 5.3 Flash (batch).

Full code generation ranking, 9 models →

general

Content Writing

  1. 1. DeepSeek: DeepSeek V4 Flash 042317%
  2. 2. DeepSeek: DeepSeek V4.1 Flash (batch)16%
  3. 3. OpenAI: GPT-6 Luna (batch)7%

DeepSeek: DeepSeek V4 Flash 0423 leads with 17% of task usage, 0 points ahead of DeepSeek: DeepSeek V4.1 Flash (batch).

Full content writing ranking, 8 models →

code

Debugging

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)33%
  2. 2. Z.ai: GLM 5.3 Flash (batch)10%
  3. 3. Crouton Alpha7%

DeepSeek: DeepSeek V4.1 Flash (batch) leads with 33% of task usage, 23 points ahead of Z.ai: GLM 5.3 Flash (batch).

Full debugging ranking, 9 models →

general

Translation

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)19%
  2. 2. Meta: Llama 3.1 8B Instruct8%
  3. 3. OpenAI: GPT-6 Luna (batch)7%

DeepSeek: DeepSeek V4.1 Flash (batch) leads with 19% of task usage, 11 points ahead of Meta: Llama 3.1 8B Instruct.

Full translation ranking, 9 models →

general

Summarization

  1. 1. StepFun: Step 5 Preview12%
  2. 2. OpenAI: GPT-6 Luna (batch)8%
  3. 3. DeepSeek: DeepSeek V4 Flash 04237%

StepFun: Step 5 Preview leads with 12% of task usage, 5 points ahead of OpenAI: GPT-6 Luna (batch).

Full summarization ranking, 9 models →

general

Customer Support

  1. 1. OpenAI: GPT-6 Luna (batch)27%
  2. 2. DeepSeek: DeepSeek V4.1 Flash (batch)11%
  3. 3. Google: Gemini 2.5 Flash8%

OpenAI: GPT-6 Luna (batch) leads with 27% of task usage, 16 points ahead of DeepSeek: DeepSeek V4.1 Flash (batch).

Full customer support ranking, 9 models →

general

Roleplay & Fiction

  1. 1. DeepSeek: DeepSeek V4 Flash 042324%
  2. 2. DeepSeek: DeepSeek V4.1 Flash (batch)24%
  3. 3. Mistral: Mistral Nemo5%

DeepSeek: DeepSeek V4 Flash 0423 leads with 24% of task usage, 0 points ahead of DeepSeek: DeepSeek V4.1 Flash (batch).

Full roleplay & fiction ranking, 8 models →

general

Q&A & Knowledge

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)15%
  2. 2. OpenAI: GPT-6 Luna (batch)8%
  3. 3. StepFun: Step 5 Preview4%

DeepSeek: DeepSeek V4.1 Flash (batch) leads with 15% of task usage, 7 points ahead of OpenAI: GPT-6 Luna (batch).

Full q&a & knowledge ranking, 5 models →

general

Classification

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)8%
  2. 2. OpenAI: GPT-6 Luna (batch)6%
  3. 3. Z.ai: GLM 5.3 Flash (batch)4%

DeepSeek: DeepSeek V4.1 Flash (batch) leads with 8% of task usage, 2 points ahead of OpenAI: GPT-6 Luna (batch).

Full classification ranking, 6 models →

Rankings by price and specialty

Comparing a 7B model against a flagship on one scale is meaningless. Each of these is a self-contained "best model" claim within a comparable price tier or use case.

By price

Under $0.50/1M

  1. 1. inclusionAI: Ling 3.1 FlashQuality 41.1
  2. 2. Anthropic: Claude Haiku 5.5 (batch)Quality 36.5
  3. 3. Z.ai: GLM 5.3 (batch)Quality 31.7
Full ranking, 33 models →

$0.50-2/1M

  1. 1. Tencent: Hy4Quality 34.6
  2. 2. Google: Gemini 3.8 Flash (batch)Quality 30.8
  3. 3. Xiaomi: MiMo-V2.6-ProQuality 29.9
Full ranking, 34 models →

$2-10/1M

  1. 1. Anthropic: Claude Opus 5.5 (batch)Quality 59.1
  2. 2. OpenAI: GPT-6.1 Sol (batch)Quality 54.6
  3. 3. Anthropic: Claude Sonnet 5.5 (batch)Quality 54.5
Full ranking, 22 models →

By specialty

Best for coding

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)53.7%
  2. 2. Z.ai: GLM 5.3 Flash (batch)12.8%
  3. 3. OpenAI: GPT-6 Luna (batch)6.0%
Full ranking, 23 models →

Best for agentic work

  1. 1. DeepSeek: DeepSeek V4.1 Flash (batch)44.3%
  2. 2. Z.ai: GLM 5.3 Flash (batch)24.0%
  3. 3. OpenAI: GPT-6 Luna (batch)10.7%
Full ranking, 15 models →

Best value

  1. 1. OpenAI: gpt-oss-120b (batch)692.5 pts/$
  2. 2. inclusionAI: Ling 3.1 Flash608.9 pts/$
  3. 3. Upstage: Solar Mini 4275.4 pts/$
Full ranking, 96 models →

Daily snapshots

Every published day gets its own permanent, unchanging permalink - a verifiable point in time for citations, not a page that silently drifts.

November 6, 2026November 5, 2026November 4, 2026November 3, 2026November 2, 2026November 1, 2026October 31, 2026October 30, 2026October 29, 2026October 28, 2026October 27, 2026October 26, 2026October 25, 2026October 24, 2026October 23, 2026October 22, 2026October 21, 2026October 20, 2026October 19, 2026October 18, 2026October 17, 2026October 16, 2026October 15, 2026October 14, 2026October 13, 2026October 12, 2026October 11, 2026October 10, 2026October 9, 2026October 8, 2026October 7, 2026October 6, 2026October 5, 2026October 4, 2026October 3, 2026October 2, 2026October 1, 2026September 30, 2026September 29, 2026September 28, 2026September 27, 2026September 26, 2026September 25, 2026September 24, 2026September 23, 2026September 22, 2026September 21, 2026September 20, 2026September 19, 2026September 18, 2026September 17, 2026September 16, 2026September 15, 2026September 14, 2026September 13, 2026September 12, 2026September 11, 2026September 10, 2026September 9, 2026September 8, 2026September 7, 2026September 6, 2026September 5, 2026September 4, 2026September 3, 2026September 2, 2026September 1, 2026August 31, 2026August 30, 2026August 29, 2026August 28, 2026August 27, 2026August 26, 2026August 25, 2026August 24, 2026August 23, 2026August 22, 2026August 21, 2026August 20, 2026August 19, 2026August 18, 2026August 17, 2026August 16, 2026August 15, 2026August 14, 2026August 13, 2026August 12, 2026August 11, 2026August 10, 2026August 9, 2026August 8, 2026August 7, 2026August 6, 2026August 5, 2026August 4, 2026August 3, 2026August 2, 2026August 1, 2026July 31, 2026July 30, 2026July 29, 2026July 28, 2026July 27, 2026July 26, 2026July 25, 2026July 24, 2026July 23, 2026July 22, 2026July 21, 2026July 20, 2026July 19, 2026July 18, 2026July 17, 2026July 16, 2026July 15, 2026July 14, 2026July 13, 2026July 12, 2026July 11, 2026July 10, 2026July 9, 2026July 8, 2026July 7, 2026July 6, 2026July 5, 2026

How this is calculated

No composite index. No invented weights. Two plain, independently-checkable numbers, each plotted as a time series instead of a single snapshot:

Quality per dollar

Artificial Analysis Intelligence Index ÷ blended price per 1M tokens (75% prompt / 25% completion). A plain division, recomputed from that day's real pricing and benchmark snapshot.

Usage share

Real token volume through OpenRouter, as a share of that day's total - no sampling, no estimation.

We tried composite indices three times this year (a weighted-sum composite, a weighted-geometric-mean composite, and a usage-vs-quality divergence index) and killed all three after finding real defensibility problems - see the full methodology page for that history. The lesson: publish real numbers over time, not invented ones.

Source: OpenRouter (openrouter.ai/rankings), as of November 7, 2026.

Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).

Methodology: www.smophy.ai/benchmark/methodology

Token counts originate from each provider's own tokenizer and are not directly comparable across providers.

Journalist or researcher? Pre-computed citable stats and CSV/JSON exports →

Field Test statistics, raw results and suggested citations