1,321

The price gap between the one that does the job and the best.

Updated on November 22, 2026.

Which LLM for which task at what price?

The right AI for the job.

For every task, three models: the one that can do the job, the best value, and the best regardless of price. With what each costs a month, for the volume you choose.

WoBench ratings are based on human evaluations.

310 models 10 tasks 9 public sources no sponsor in the verdicts
The best
Verdict of the day 1 of 10
Claude Fable 5 Anthropic
at the bar 1,495.5 to 1,512.7 · 1st of 131 best

Preferred 6 times out of 10 to the one that does the job

$7.50 a month

1,321 times the price of the one that does the job, which costs $0.00568 a month.

At the Small volume: 100 runs per month, of 1,500 tokens in and 1,200 out. redo the calculation

Prices are those published by the providers and can change without notice. Our verdicts guide a choice, they do not guarantee it: test the model on your own cases and validate your calculation, economics included. Our method

1,321 The price gap between the one that does the job and the best.

On Write, write texts, emails and editorial content. The, the cheapest that does the job costs $0.00568 a month, the best $7.50. 131 models measured, 61 meet the bar.

Free right now
Without paying, for now

We do not know how long these offers last, or what limits apply.

All free models
Can do the job
the economical choice that covers the need
DeepSeek
1,400.8 to 1,415.9 · bar 1,400.4
$0.00568 a month,
at 100 runs
The cheapest of the models whose range touches the bar.
Best value
the best balance between performance and cost
Z.ai
1,447.4 to 1,461.5 · bar 1,400.4
$0.0996 a month,
at 100 runs
99% cheaper than The best, 49.7 measure points less.
The best
the best performing model, regardless of price
Anthropic
1,495.5 to 1,512.7 · bar 1,400.4
$7.50 a month,
at 100 runs
1st of the 131 models this task judges. 1321 times the price of Can do the job, for 95.8 measure points more.

Scores from November 19, 2026, prices from November 22, 2026.

Small volume. The measure details and the full model list are on the task page. WoBench ratings are based on human evaluations.

See what they cost, from the cheapest to the most expensive.

One dot per model and per task. the cheapest that does the job best value the best free right now (in its zone, position follows the score, unrelated to any price) The bigger the dot, the higher the score. Compared with the entry price, scale fitted to wide gaps

61 models placed on Write, at the Small volume.

Explore the full ranking, task by task

Free right now read on November 22, 2026, 3:03 am

What costs nothing today, and what it is worth.

6 tasks of 10 have a free model that clears the bar.

Providers sometimes give access to a model for free. We show it apart, without letting it compete, because a free price says nothing about the service delivered. Each cell answers a single question: on this task, does this model reach the bar? Same measures, same bars as the verdicts.

Free models by task
Free model Sort Web Summary Synthesis RAG Write Code Web Dev Agent Images Elsewhere, paid
Gemma 4 26B A4B
Google AI Studio since September 30, 2026 data: not checked See the model page
$0.05 / $0.19
paid at 14 providers, per million tokens
Gemma 4 31B
Google AI Studio since September 30, 2026 data: not checked See the model page
$0.10 / $0.10
paid at 10 providers, per million tokens
Nemotron 3 Ultra
Nvidia since September 30, 2026 data: not checked See the model page
$0.50 / $2.15
paid at 4 providers, per million tokens
Clears the bar Below the bar Not measured Requirement not met

Choose a cell to read its detail here.

8 other free models clear the bar on no task

Read before choosing

Free offers and prices are those published by providers, and may change or stop without notice. Our verdicts guide a choice, they do not guarantee it: test the model on your own cases and check your numbers, including the economics, planning for what the same work will cost on the day the offer stops.

Duration

Until when?

No offer announces its end. The 7 we have seen stop lasted from 4 to 39 days.

Limits

How many requests?

Free access may be capped in requests or throughput. Our sources do not publish these caps.

Your data

Who reads your requests?

Free access may allow the provider to reuse what you send. Read its terms before sending customer data.

Tracked since September 30, 2026, and still ongoing.

See all free offers and their history

The market on one screen, 10 tasks, 29 verdicts.

One row per task, the three verdicts, at the Small volume. Each row opens its task.

Task Can do the job Best value The best Gap
Sort and extract Sort, tag and extract structured data at high volume.
DeepSeek V4 Flash 0423 DeepSeek $0.0163 ↓ -72%
GLM 5.3 Flash Z.ai $0.092
Claude Fable 5 Anthropic $12.80
×785
Search the web Search and summarise information on the web.
Claude Opus 4.6 Anthropic $5.00
GPT-5.6 Sol OpenAI $7.20
×7.6
Summarise Summarise a short to medium text, article or report.
DeepSeek V4 Flash 0423 DeepSeek $0.014 ↓ -72%
GLM 5.3 Flash Z.ai $0.08
Claude Fable 5 Anthropic $12.00
×859
Synthesize Summarise or analyse 50K to 1M tokens.
DeepSeek V4 Flash 0423 DeepSeek $0.0303 ↓ -72%
GLM 5.3 Flash Z.ai $0.16
Claude Fable 5 Anthropic $22.00
×726
RAG Answer questions from the documents you supply.
GPT-5.6 Luna OpenAI $1.96
Claude Sonnet 4.6 Anthropic $17.64
Claude Opus 5 Anthropic $44.40
×22.7
Write Write texts, emails and editorial content.
DeepSeek V4 Flash 0423 DeepSeek $0.00568 ↓ -72%
GLM 5.2 Z.ai $0.0996
Claude Fable 5 Anthropic $7.50
×1,321
Code Write, fix or explain code.
DeepSeek V4 Flash 0423 DeepSeek $0.0786 ↓ -72%
GLM 5.3 Flash Z.ai $0.50
Claude Fable 5 Anthropic $78.00
×992
Web Dev A whole website or web app from a brief, judged on the result.
DeepSeek V4 Flash 0423 DeepSeek $0.0364 ↓ -72%
GLM 5.3 Z.ai $0.83
Claude Opus 5 Anthropic $10.50
×288
Agent Long loops with tool calls and recovery from errors.
DeepSeek V4 Pro 0423 DeepSeek $1.11 ↓ -26%
no model holds this verdict
Claude Opus 4.7 Anthropic $13.05
×11.8
Analyse images Describe, explain or answer a question about a photo, a screenshot or a chart.
Gemma 4 26B A4B Google $0.049
GLM 5.3 Flash Z.ai $0.07
Claude Fable 5 Anthropic $11.00
×224

What changed since the last update.

updated on November 22, 2026 Every morning, after the sources are read. Compared with the state of November 21, 2026. Everything that changed

Recommended models that change 0

Nothing has changed since November 21, 2026.

Price drops 6

6 of 9, the other 3

Latest models added to the catalogue 3

No model scored since the previous publication. Here are the latest three added to the catalogue.

  • Nova 2 Lite too little data Amazon, added to the catalogue on September 3, 2026. Not enough votes to place it.
  • Nova Lite 1.0 too little data Amazon, added to the catalogue on September 3, 2026. Not enough votes to place it.
  • Nova Micro 1.0 does not do the job Amazon, added to the catalogue on September 3, 2026. On none of our 10 tasks.
×1,321
Figure of the day

The cost gap between the best and the cheapest that does the job, on Write.

The cheapest that does the job, DeepSeek V4 Flash 0423, costs $0.00568 a month. The best, Claude Fable 5, costs $7.50. At 100 runs per month, for a task both can do.

Prices: OpenRouter (providers), collected on November 22, 2026.

Ratings: LMArena, creative writing leaderboard.

Open Write
Your WoBench

Yesterday's right call, still right today?

Follow the models you use: on the task you care about, each one says where it stands, and what it costs compared with the one that does the job. The rest of the time, we stay quiet.

Free. Leave whenever you like, we delete everything.

Where these numbers come from European Central Bank Epoch AI LiteLLM LMArena models.dev OpenRouter (providers and per provider prices) OpenRouter (models) tau²-bench (Sierra Research) Vectara (hallucination leaderboard) All sources, licences and dates Prices are collected once a day, at 3:00 am (Paris time).

These figures come from the public sources cited and reflect the last import, not the state at this second. Prices are those published by the providers and can change without notice: check them there before committing.

Costs are computed on a reference volume per task, shown on the page of each task: your usage will give another amount. Our method