MDL.01
DeepSeek
DeepSeek's open-weight models for visual documents, coding and model routing: the current release, dated specifications, API price comparison and migration guidance.
Open pageTechnology
AI model families.
Each on its merits.
We are not a reseller for any of these and receive nothing for recommending one. What follows is what each family is genuinely good at, what it costs, where its weights and data live, and the conditions under which we would pick something else.
- Model-agnostic
- Benchmarked per workload
Introduction transcript. Qubi: There's no best model, only your work and the limits around it. Your limits come first: where data may live, the licence, the budget. Then a short list is tested on your own tasks, for quality, cost and speed. Whichever fits is pinned behind an interface, and tested again when models change.
New decision model
Meet Jev from TypeSafe AI.
Jev powers fast, typed decisions inside software. It adds a different role to the model catalogue: selecting tools, routing requests and assessing a bounded condition. A generative model can still produce the answer or code around it.
Explore our model-router design and the published benchmark comparison to see where that decision layer may fit.
The selection is not a ranking
There is no best model. There is a workload, a set of constraints, and a family that fits them better than the others do. Capability is rarely the binding constraint — most production tasks sit comfortably inside the reach of six or seven families, and the decision is settled by where the data is permitted to live, what latency the interface can absorb, and what a request costs at your projected volume.
That is why these pages are organised by family rather than by leaderboard position. A leaderboard tells you which model won a benchmark somebody else designed. It does not tell you that your extraction task is handled identically by a model a twentieth of the price, which is the finding that changes a budget.
Two structural questions cut across all of them, and they are worth settling before you read a single capability claim. First: open weights or hosted API? That decides whether you can run the model on hardware you control, which in turn decides whether a whole class of data-residency conversation happens at all. Second: who trained it, and who operates the endpoint? Those are different organisations for open-weight models, and the distinction matters more than most procurement processes currently reflect.
The families at a glance
Compare your options
| Criterion | Weights | Runs on your hardware | Typical role |
|---|---|---|---|
| Anthropic Claude | Closed | No | Long-context reasoning, agentic tool use |
| OpenAI GPT | Closed | No | Broad general capability, computer use, mature ecosystem |
| Google Gemini | Closed | No | Very long context, native multimodal |
| DeepSeek | Open | Yes | Frontier-adjacent reasoning at open-weight cost |
| Meta Llama & Muse | Open (community licence) | Yes | The default self-hosted workhorse |
| Mistral | Open (some Apache-2.0) | Yes | EU-domiciled provider, some Apache-2.0 licences |
| Qwen | Open | Yes | Widest size range, strong multilingual |
| Z.ai GLM | Open for much of the range (licence varies) | Yes | Self-hosted coding and agent work |
| Kimi | Open for some releases (licence varies) | Yes | Very large open releases; newest flagships API-first |
| Grok | Closed | No | A further hosted frontier candidate |
| MiniMax | Open (community licence) | Yes | Low-cost, high-volume work |
| Small models (1B–14B) | Mostly open | Yes | High-volume narrow tasks at low unit cost |
NEW RELEASES
New releases, across every family.
Models the public catalogues listed in the last few weeks, newest first. Each has its own page with specifications, independent ratings and listed prices.
Mistral · 29 Oct 2026 · open weights
Mistral Large 4
- Intelligence
- 26.0
- Context
- 1.05M
- In / out per M
- $0.700 / $2.10
OpenAI GPT · 26 Oct 2026
GPT-6.1 Terra
- Intelligence
- 51.0
- Context
- 1.05M
- In / out per M
- $1.00 / $5.00
Kimi · 23 Oct 2026
Kimi K3.1 Pro
- Intelligence
- 46.0
- Context
- 1.05M
- In / out per M
- $2.00 / $10.00
MiniMax · 19 Oct 2026 · open weights
MiniMax M3.1 Flash
- Intelligence
- 28.9
- Context
- 1.05M
- In / out per M
- $0.075 / $0.300
DeepSeek · 17 Oct 2026 · open weights
DeepSeek V4.1 Flash 1017
- Intelligence
- 32.5
- Context
- 1.05M
- In / out per M
- $0.150 / $0.600
Meta · 15 Oct 2026 · open weights
Muse Spark Open
- Intelligence
- 24.7
- Context
- 1.05M
- In / out per M
- $1.25 / $4.25
Meta · 12 Oct 2026
Muse Wonder
- Intelligence
- 43.3
- Context
- 1.05M
- In / out per M
- $4.00 / $20.00
Claude · 7 Oct 2026
Claude Haiku 5.5
- Intelligence
- 36.2
- Context
- 1M
- In / out per M
- $0.100 / $0.500
CURRENT RELEASES
What each family leads with now.
Checked against the public model catalogues every week. A new release is reviewed against the provider's own announcement before a page names it.
| Criterion | Leads with | Released | Page reviewed |
|---|---|---|---|
| Claude | Claude Fable 5.1 | 1 September 2026 | 19 September 2026 |
| OpenAI GPT | GPT-6 Astra | 3 September 2026 | 5 September 2026 |
| Gemini | The family as a range | — | 1 September 2026 |
| DeepSeek | DeepSeek V4.1-Flash | 10 September 2026 | 19 September 2026 |
| Meta | The family as a range | — | 1 September 2026 |
| Mistral | The family as a range | — | 1 September 2026 |
| Qwen | The family as a range | — | 1 September 2026 |
| Z.ai GLM | The family as a range | — | 30 October 2026 |
| Kimi | The family as a range | — | 30 October 2026 |
| Grok | The family as a range | — | 30 October 2026 |
| MiniMax | The family as a range | — | 30 October 2026 |
INDEPENDENT RATINGS
How the families compare.
Third-party scores and listed prices, refreshed every week. Useful for building a shortlist; your own evaluation decides.
Composite index across their reasoning, knowledge, maths and coding evaluations · Higher is better · Linear scale from 0 to 66.0
- Gemini 4 Carbon
- 59.4
- Claude Opus 5.5
- 58.4
- GPT-6 Astra
- 56.9
- Muse Wonder Full
- 49.7
- Kimi K3.1 Pro
- 46.0
- Grok 4.7
- 35.0
- Qwen3.8 Max
- 33.6
- DeepSeek V4.1 Flash 1017
- 32.5
The highest-rated model of each family we cover. One composite score cannot show cost, latency, data residency or how a model handles your task.
Source: Artificial Analysis. Checked 30 October 2026. Measured independently by Artificial Analysis, not by QQuantum.ai or the providers.
EVERY RATED MODEL
The top-rated models, side by side.
The highest-rated current models across every family we cover. Sort by any column; select a model for its own page.
| Claude Opus 5.5 | Claude | 22 Sep 2026 | 58.4 | 98 | 1M | $4.00 | $20.00 |
|---|---|---|---|---|---|---|---|
| GPT-6 Astra | OpenAI GPT | 3 Sep 2026 | 56.9 | 88 | 1.05M | $10.00 | $50.00 |
| GPT-6.1 Sol | OpenAI GPT | 29 Sep 2026 | 54.5 | 84 | 1.05M | $2.00 | $10.00 |
| Claude Sonnet 5.5 | Claude | 28 Sep 2026 | 54.2 | 135 | 1M | $2.00 | $10.00 |
| GPT-6.1 Terra | OpenAI GPT | 26 Oct 2026 | 51.0 | 91 | 1.05M | $1.00 | $5.00 |
| Gemini 4 Argon | Gemini | 30 Sep 2026 | 50.4 | 80 | 1.05M | $2.00 | $10.00 |
| Claude Fable 5.1 | Claude | 1 Sep 2026 | 49.0 | 72 | 1M | $10.00 | $50.00 |
| Kimi K3.1 Pro | Kimi | 23 Oct 2026 | 46.0 | 30 | 1.05M | $2.00 | $10.00 |
| Muse Wonder | Meta | 12 Oct 2026 | 43.3 | 150 | 1.05M | $4.00 | $20.00 |
| Muse Spark 1.3 | Meta | 2 Sep 2026 | 36.4 | — | 1.05M | $1.25 | $4.25 |
| Claude Haiku 5.5 | Claude | 7 Oct 2026 | 36.2 | 235 | 1M | $0.100 | $0.500 |
| Grok 4.7 | Grok | 21 Sep 2026 | 34.3 | 82 | 500k | $2.00 | $6.00 |
| Qwen3.8 Max (0902) | Qwen | 2 Sep 2026 | 33.6 | 38 | 1M | $2.00 | $6.00 |
| DeepSeek V4.1 Flash 1017 open | DeepSeek | 17 Oct 2026 | 32.5 | 159 | 1.05M | $0.150 | $0.600 |
| GLM 5.3 open | Z.ai GLM | 18 Aug 2026 | 31.7 | 76 | 1.05M | $0.080 | $3.60 |
| Gemini 3.8 Flash | Gemini | 2 Sep 2026 | 30.5 | 126 | 1.05M | $0.825 | $4.13 |
| MiniMax M3.1 Flash open | MiniMax | 19 Oct 2026 | 28.9 | 111 | 1.05M | $0.075 | $0.300 |
| GPT-6 Luna | OpenAI GPT | 22 Sep 2026 | 27.8 | 116 | 1.05M | $0.100 | $0.500 |
| GLM 5.3 Flash open | Z.ai GLM | 26 Aug 2026 | 27.8 | 66 | 1.05M | $0.018 | $0.140 |
| Mistral Large 4 open | Mistral | 29 Oct 2026 | 26.0 | 169 | 1.05M | $0.700 | $2.10 |
| Muse Spark Open open | Meta | 15 Oct 2026 | 24.7 | 140 | 1.05M | $1.25 | $4.25 |
| DeepSeek V4 Pro 0813 open | DeepSeek | 13 Aug 2026 | 23.1 | 77 | 1.05M | $0.660 | $1.98 |
| DeepSeek V4 Pro 0423 open | DeepSeek | 13 Aug 2026 | 23.1 | 77 | 1.05M | $0.948 | $1.90 |
| Qwen3.8 27B open | Qwen | 14 Aug 2026 | 20.1 | 39 | 1M | $0.150 | $1.88 |
| MiniMax M3 open | MiniMax | 1 Jun 2026 | 18.7 | 110 | 1.05M | $0.300 | $1.20 |
| Gemini 3.5 Flash Lite | Gemini | 21 Jul 2026 | 11.7 | 399 | 1.05M | $0.300 | $2.50 |
| Mistral Small 4 open | Mistral | 16 Mar 2026 | 4.3 | 170 | 262k | $0.150 | $0.600 |
Intelligence and coding: Artificial Analysis indices at the model's best reasoning setting, and its measured output speed. Context and prices: the OpenRouter catalogue list price in USD per million tokens. Released: the provider's date where Artificial Analysis records one, otherwise the date the catalogue first listed it. Refreshed weekly; not measured by QQuantum.ai. Select a heading to sort.
IMAGE, VIDEO AND VOICE
Generative media, ranked weekly.
Image, video and voice models ranked by blind preference votes, each with its own page.
123 models ranked
Image generation
Leading now: GPT Image 2.5 Sunburst (max) (OpenAI), Elo 1222.
See the leaderboard →37 models ranked
Video generation
Leading now: HeyGen Video 1 (HeyGen), Elo 1143.
See the leaderboard →100 models ranked
Voice (text to speech)
Leading now: Eleven v4 Turbo (ElevenLabs), Elo 1336.
See the leaderboard →Every model page
-
-
MDL.02
Claude
Anthropic Claude for enterprise coding and long-running agents: the current release, dated benchmark evidence, access paths and evaluation guidance.
Open page -
MDL.03
OpenAI GPT Integration
Where GPT models fit, what computer use and a million-token window make buildable, and the trade-offs against other frontier and open options.
Open page -
MDL.04
Gemini for Enterprise
Long-context and multimodal strengths, Vertex deployment, and where Gemini earns its place in a stack.
Open page -
MDL.05
Meta Llama & Muse
Meta's models, Llama and the newer Muse line — licence terms, self-hosting, ecosystem maturity and realistic serving costs.
Open page -
MDL.06
Mistral: European Open Weights
European open-weight models, Apache-licensed options and the data-residency case for using them.
Open page -
MDL.07
Qwen for Multilingual Work
Alibaba Qwen — broad model sizes, strong multilingual coverage, and the same provenance questions as any non-EU open model.
Open page -
MDL.08
Z.ai GLM
Z.ai's GLM models: open-weight and API releases for coding and agent work, licences that differ by release, and the provenance review a Chinese provider needs.
Open page -
MDL.09
Kimi by Moonshot AI
Moonshot AI's Kimi models: very large open-weight releases, API-only flagships, what self-hosting them really takes and how to evaluate them.
Open page -
MDL.10
Grok
SpaceXAI's Grok models (formerly xAI): a closed frontier line by API, where it fits, and what to check before relying on it.
Open page -
MDL.11
MiniMax
MiniMax's open-weight language models: low listed prices, a community licence, and how to test them for high-volume work.
Open page -
MDL.12
Image Generation Models
Every image generation and editing model ranked by blind preference votes, refreshed weekly, with a page per model and how we evaluate them for real briefs.
Open page -
MDL.13
Video Generation Models
Text-to-video and image-to-video models ranked by blind preference votes, refreshed weekly, with a page per model and the checks before production use.
Open page -
MDL.14
Voice Models (Text to Speech)
Text-to-speech models ranked by blind listening votes, refreshed weekly, with a page per model and what to test before a voice goes live.
Open page -
MDL.15
Open-Weight Models
What running your own models genuinely costs, when the control is worth the capability gap, and how to decide.
Open page -
MDL.16
Small Language Models
The 1B–14B class that handles most production tasks at a fraction of frontier cost — and where it stops being enough.
Open page -
MDL.17
Embedding & Reranking Models
The models that decide retrieval quality, why they matter more than the generator, and how to evaluate them on your corpus.
Open page
How we actually decide
SELECTION PROCEDURE
-
01Constraints
residency · licence · budget
-
02Candidate set
usually 3–4 families
-
03Eval harness
your data, your task
-
04Cost & latency
at projected volume
-
05Pin a version
behind an interface
-
06 · Branch / escalationRe-run on change
deprecation is routine
Solid: selection pathDashed: the loop that runs for the life of the system
Questions
Questions about Model Families.
Do you have a preferred model?
No, and we have no commercial reason to. We hold no reseller or partner agreement with any model provider, so the recommendation is not worth anything to us beyond being right. In practice we run three or four candidates against your task and the result is frequently that two of them would work and the choice comes down to price or an infrastructure relationship you already have.
How often does the answer change?
The shortlist changes roughly every two quarters; the decision procedure does not. That is why we build systems where the model sits behind an interface — swapping a family should be an evaluation run and a config change, not a rewrite. Systems whose prompts are entangled with one model version do not survive a deprecation cycle.
Can we mix families in one system?
Usually you should. Routing straightforward requests to a small or open model and reserving a frontier model for the genuinely hard minority is the single largest cost lever available, and it typically costs nothing in quality when the routing is measured rather than guessed. See hybrid model routing for how the gate is built.
What if our data cannot leave our infrastructure?
Then the closed families are out and the decision is made among the open-weight ones, running on hardware you control. That is a real capability trade-off and we would rather state it than paper over it — though the gap has narrowed enough that for many production tasks it no longer decides anything.
Further reading
What changed this September?
Our September 2026 AI briefing connects current releases to customer decisions, including DeepSeek for visual workloads and Claude for longer coding and knowledge tasks.
