Open Models have crossed a threshold
📰 LangChain Blog
Open models like GLM-5 and MiniMax M2.7 now match closed frontier models on core agent tasks at a fraction of the cost and latency
Action Steps
- Evaluate open models like GLM-5 and MiniMax M2.7 for core agent tasks
- Compare the cost and latency of open models to closed frontier models
- Consider using open models for production workflows to reduce costs and improve response times
- Explore specialized inference infrastructure providers like Groq, Fireworks, and Baseten to optimize latency and throughput
Who Needs to Know This
Developers and data scientists on a team can benefit from using open models for agent tasks, as they offer a viable option for reducing costs and latency while maintaining performance
Key Insight
💡 Open models offer a level of consistency and predictability that makes real-world workflows more viable, while reducing costs and latency
Share This
🚀 Open models now match closed frontier models on core agent tasks at a fraction of the cost and latency! 💸
Key Takeaways
Open models like GLM-5 and MiniMax M2.7 now match closed frontier models on core agent tasks at a fraction of the cost and latency
Full Article
Published Time: 2026-04-02T17:51:54.000Z
# Open Models have crossed a threshold
[Skip to content](https://blog.langchain.com/open-models-have-crossed-a-threshold/#main)
[](https://blog.langchain.com/)
* [Website](https://www.langchain.com/)
* [Docs](https://docs.langchain.com/)
* [Case Studies](https://blog.langchain.com/tag/case-studies/)
* [Harrison's In the Loop Series](https://blog.langchain.com/tag/in-the-loop/)
* [Try LangSmith](https://smith.langchain.com/)
[Sign in](https://blog.langchain.com/open-models-have-crossed-a-threshold/#/portal/signin)[Subscribe](https://blog.langchain.com/open-models-have-crossed-a-threshold/#/portal/signup)
# Open Models have crossed a threshold
6 min read Apr 2, 2026
💡
****TL;DR:**** Open models like GLM-5 and MiniMax M2.7 now match closed frontier models on core agent tasks — file operations, tool use, and instruction following — at a fraction of the cost and latency. Here's what our evals show and how to start using them in Deep Agents.
Over the past few weeks, we’ve been running open weight Large Language Models through [Deep Agents](https://github.com/langchain-ai/deepagents?ref=blog.langchain.com) harness evaluations, and the initial results show they are a viable option to use instead of, and alongside, closed frontier models. GLM-5 ([z.ai](http://z.ai/?ref=blog.langchain.com)) and [MiniMax](https://www.minimax.io/models/text/m27?ref=blog.langchain.com) M2.7 each score similarly to closed frontier models on core agent tasks such as file operations, tool use, and instruction following.
This isn’t surprising if you’ve been following open model progress via the large set of open benchmarks such as [SWE-Rebench](https://swe-rebench.com/?ref=blog.langchain.com) and [Terminal Bench 2.0](https://www.tbench.ai/leaderboard/terminal-bench/2.0?ref=blog.langchain.com). Tool calling is reliable and instruction following is consistent. For developers deploying agents in production, open models now offer a level of consistency and predictability that makes real-world workflows much more viable.
## Why open models
When exploring open models, builders and customers tend to focus on a few key factors: **cost, latency,** and **task performance**.
In the limit, it would be great to use the smartest frontier model at the highest reasoning level for every task. In practice, two constraints make that unworkable: cost and latency. Closed frontier models can run 8–10x more expensive for high-throughput workloads, and they're often too slow for the response times users expect in interactive products.
| Model | Type | Input ($/M tokens) | Output ($/M tokens) |
| --- | --- | --- | --- |
| Claude Opus 4.6 (Anthropic) | Closed | $5.00 | $25.00 |
| Claude Sonnet 4.6 (Anthropic) | Closed | $3.00 | $15.00 |
| GPT-5.4 (OpenAI) | Closed | $2.50 | $15.00 |
| GLM-5 (Baseten) | Open | $0.95 | $3.15 |
| MiniMax M2.7 (OpenRouter) | Open | $0.30 | $1.20 |
_To put the pricing in context: an application outputting 10M tokens/day costs roughly $250/day on Opus 4.6 versus ~$12/day for MiniMax M2.7. That's about a $87k annual difference._
Open models tend to be smaller than closed frontier models, and can be accelerated on specialized inference infrastructure — providers like [Groq](https://groq.com/?ref=blog.langchain.com), [Fireworks](https://fireworks.ai/?ref=blog.langchain.com), and [Baseten](https://www.baseten.co/?ref=blog.langchain.com) optimize for latency and throughput far beyond what most teams could achieve on their own. [OpenRouter data](https://openrouter.ai/z-ai/glm-5/performance?ref=blog.langchain.com) show GLM-5 on Baseten averaging 0.65s latency and 70 tokens/second, compared to 2.56s and 34 tokens/second for Claude Opus 4.6. For latency-sensitive produc
# Open Models have crossed a threshold
[Skip to content](https://blog.langchain.com/open-models-have-crossed-a-threshold/#main)
[](https://blog.langchain.com/)
* [Website](https://www.langchain.com/)
* [Docs](https://docs.langchain.com/)
* [Case Studies](https://blog.langchain.com/tag/case-studies/)
* [Harrison's In the Loop Series](https://blog.langchain.com/tag/in-the-loop/)
* [Try LangSmith](https://smith.langchain.com/)
[Sign in](https://blog.langchain.com/open-models-have-crossed-a-threshold/#/portal/signin)[Subscribe](https://blog.langchain.com/open-models-have-crossed-a-threshold/#/portal/signup)
# Open Models have crossed a threshold
6 min read Apr 2, 2026
💡
****TL;DR:**** Open models like GLM-5 and MiniMax M2.7 now match closed frontier models on core agent tasks — file operations, tool use, and instruction following — at a fraction of the cost and latency. Here's what our evals show and how to start using them in Deep Agents.
Over the past few weeks, we’ve been running open weight Large Language Models through [Deep Agents](https://github.com/langchain-ai/deepagents?ref=blog.langchain.com) harness evaluations, and the initial results show they are a viable option to use instead of, and alongside, closed frontier models. GLM-5 ([z.ai](http://z.ai/?ref=blog.langchain.com)) and [MiniMax](https://www.minimax.io/models/text/m27?ref=blog.langchain.com) M2.7 each score similarly to closed frontier models on core agent tasks such as file operations, tool use, and instruction following.
This isn’t surprising if you’ve been following open model progress via the large set of open benchmarks such as [SWE-Rebench](https://swe-rebench.com/?ref=blog.langchain.com) and [Terminal Bench 2.0](https://www.tbench.ai/leaderboard/terminal-bench/2.0?ref=blog.langchain.com). Tool calling is reliable and instruction following is consistent. For developers deploying agents in production, open models now offer a level of consistency and predictability that makes real-world workflows much more viable.
## Why open models
When exploring open models, builders and customers tend to focus on a few key factors: **cost, latency,** and **task performance**.
In the limit, it would be great to use the smartest frontier model at the highest reasoning level for every task. In practice, two constraints make that unworkable: cost and latency. Closed frontier models can run 8–10x more expensive for high-throughput workloads, and they're often too slow for the response times users expect in interactive products.
| Model | Type | Input ($/M tokens) | Output ($/M tokens) |
| --- | --- | --- | --- |
| Claude Opus 4.6 (Anthropic) | Closed | $5.00 | $25.00 |
| Claude Sonnet 4.6 (Anthropic) | Closed | $3.00 | $15.00 |
| GPT-5.4 (OpenAI) | Closed | $2.50 | $15.00 |
| GLM-5 (Baseten) | Open | $0.95 | $3.15 |
| MiniMax M2.7 (OpenRouter) | Open | $0.30 | $1.20 |
_To put the pricing in context: an application outputting 10M tokens/day costs roughly $250/day on Opus 4.6 versus ~$12/day for MiniMax M2.7. That's about a $87k annual difference._
Open models tend to be smaller than closed frontier models, and can be accelerated on specialized inference infrastructure — providers like [Groq](https://groq.com/?ref=blog.langchain.com), [Fireworks](https://fireworks.ai/?ref=blog.langchain.com), and [Baseten](https://www.baseten.co/?ref=blog.langchain.com) optimize for latency and throughput far beyond what most teams could achieve on their own. [OpenRouter data](https://openrouter.ai/z-ai/glm-5/performance?ref=blog.langchain.com) show GLM-5 on Baseten averaging 0.65s latency and 70 tokens/second, compared to 2.56s and 34 tokens/second for Claude Opus 4.6. For latency-sensitive produc
DeepCamp AI