I want more models

[Elijah82236801]

🏷️ Discussion Type

Question

💬 Feature/Topic Area

Code Search and Navigation

Body

I want more models, like:

• Claude
• Copilot
• Gemini
• Groq
• Midjourney

[codemax101]

More AI Chat Models:

• Llama 4 Scout (Meta): Meta shook up the market recently by introducing an industry-leading 10-million token context window. This means you can upload entire libraries of books or massive corporate databases into the model's memory all at once.

• Qwen 3.7 Max (Alibaba): A massive disruptor from China. It operates at the frontier level for reasoning and competition-level math, but does so at a fraction of the cost of Western proprietary models.

• MAI Models (Microsoft): At Microsoft Build 2026, CEO Mustafa Suleyman unveiled seven new models under the "MAI" banner (including MAI Image-2.5 and MAI Voice-2), signaling Microsoft's move to build its own distinct AI ecosystem alongside its partnership with OpenAI.

[Kunal241207]

If you want to look outside of those five, here are the other main models worth checking-

• GPT-5.5 / 5.4: The biggest pro is the adaptive routing that seamlessly flips to a deep "thinking" mode for hard coding tasks, plus its native computer-use tool execution. The obvious con is that it's pricey to run compared to open alternatives.

• GLM-5.2: Built strictly for agentic, repository-scale coding with a massive 1-million token context window. The big pro is that its logic can easily handle multi-file diffs and web dev tasks, but the con is that it is highly token-hungry.

• Kimi (K2.6): Famous for its "Agent Swarm" tech that can orchestrate dozens of sub-agents at once for massive research and coding loops. The catch is that it can occasionally drop progress or stall on exceptionally long, hour-spanning engineering tasks.

• Gemma 4 (31B): Google's flagship open-weights model designed specifically to run locally on developer workstations. It has incredible function-calling and structured layout reasoning for its size, but it lacks the ultra-massive multi-million token context window of cloud-gated giants.

• DeepSeek (V3/R1): Its coding logic easily rivals Claude and it’s dirt cheap to use. The downside is it can be pretty verbose and rigid with formatting.

• Qwen (2.5-Coder): Incredible open-weights model for pinpointing bugs and writing clean code, but it's pretty weak at general, non-coding tasks.

• Flux.1: The best Midjourney alternative for UI mockups because it actually renders text and UI elements cleanly. The catch is it lacks Midjourney's automatic cinematic styling and needs a lot of local power.

[jogimj13-cyber]

Model for Coding

All models are on Openrouter.

Low budget

Grok 0.1 Build

xAI's promising agentic model is only $1 per every million input tokens, and $2 per every million output tokens. Strong choice if budget allows.

Deepseek V4 Pro/Flash

Both models are impressive in code - and are less than $1. Easily the one of the best models to their price

	Pro	Flash
Input	$0.435	$0.09
Output	$0.87	$0.18

Nemotron 3 Ultra

Free, and decent in coding

Gemini 2.5 Flash Lite

Input - $0.10
Output - $0.40

Medium budget

GPT 5.4 Nano

GPT 5.4 Nano is OpenAI's strongest nano model - and it's not too expensive
Input - $0.20
Output - $1.25

Trinity Large Thinking

Input - $0.50
Output - $0.80

GLM 5.2

Input - $0.95
Output - $3

Kimi K2.7 Code

MoonshotAI has a specialized model for code
Input - $0.74/ M tokens
Output - $3.50/ M tokens

High Budget

Gemini 3.5 Flash

High pricing but great quality
Input - $1.50
Output - $9

GPT 5.3 Codex

Input - $1.75
Output - $14

Very High Budget

Claude Opus 4.8

Input - $5
Output - $25

GPT 5.5

Input - $5
Output - $30

Claude Fable 5

Ridiculously good at ANY coding task. The downside? It's expensive.
Input: $10
Output: $50

Picks

Claude Fable 5

If budget allows

Deepseek V4 Flash

Very inexpensive, with good quality with it

Table

Model	Provider	Input Price (per 1M tokens)	Output Price (per 1M tokens)	Quality
Nemotron 3 Ultra	NVIDIA	Free	Free	Decent, reliable, lower-end
Deepseek V4 Flash	Deepseek	$0.09	$0.18	Fast & cheap, very good code reliability
Gemini 2.5 Flash Lite	Google	$0.10	$0.40	Solid, reasonable QC, good for cost
Deepseek V4 Pro	Deepseek	$0.435	$0.87	Excellent for price, robust
Trinity Large Thinking	Arcee-AI	$0.50	$0.80	Reliable, production-level code
GPT 5.4 Nano	OpenAI	$0.20	$1.25	Strongest nano, highly capable
Grok 0.1 Build	xAI	$1.00	$2.00	Strong code, agentic, high value
Kimi K2.7 Code	Moonshot	$0.74	$3.50	Specializes in code, modern, consistent
GLM 5.2	Z.ai	$0.95	$3.00	Good, not top tier, but consistent
Gemini 3.5 Flash	Google	$1.50	$9.00	High quality, fast, up-to-date
GPT 5.3 Codex	OpenAI	$1.75	$14.00	Strong code, trusted, but pricey
Claude Opus 4.8	Anthropic	$5.00	$25.00	Premium, near SOTA, very strong output
GPT 5.5	OpenAI	$5.00	$30.00	SOTA, enterprise-grade
Claude Fable 5	Anthropic	$10.00	$50.00	Unmatched performance on complex coding tasks; top-tier accuracy, reasoning, and reliability—best-in-class if budget permits.

[amrthesha213-netizen]

In addition to the models already mentioned, it would be great to see support for more specialized options depending on the task.

• Llama 4: A strong open-weight family that gives developers more control over deployment, especially for self-hosted or enterprise environments.
• Mistral Large: Known for good reasoning performance and competitive coding capabilities while offering another commercial alternative.
• Codestral : A model specifically optimized for software engineering tasks such as code generation, completion, and refactoring.
• DeepSeek R1: A solid choice for reasoning-heavy coding workflows and debugging, particularly when cost efficiency is important.
• Qwen2.5-Coder: Performs well on code understanding, bug fixing, and repository navigation, making it a useful option for development-focused workflows.

Rather than simply adding more models, it would also be valuable to let users choose the model based on the task they're performing. For example, one model may be better suited for code reviews, another for documentation, and another for long-context repository analysis. That kind of flexibility would make the platform useful for a much broader range of developers.

[KateCatlin]

Hey all! We hear you and we are working on it. We just released Kimi K2.7 Code to Pro/Pro+/Max plans, and are slowly rolling it out: https://github.blog/changelog/2026-06-30-kimi-k2-7-is-now-available-in-github-copilot

Give it a whirl and let us know what you think!

Community Manager Update (July 1, 2026):

Important

Effective today, July 1, 2026, Anthropic’s Claude Fable 5 is now re-enabled and generally available in GitHub Copilot.

Read more about this release in the Changelog: Claude Fable 5 is generally available for GitHub Copilot