GPT-6.2 Astra and Sol are live on TokenRouter.
Astra lists at $8/$40 per 1M tokens, 20% below GPT-6 Astra, with cached input down to $0.40.
Sol stays at $2/$10 and scores 57 on the Artificial Analysis Intelligence Index, level with GPT-6 Astra.
tokenrouter.com
1
6
359
Qwen4-Max is live on TokenRouter.
1M context with native vision, at $1.50 in and $4.50 out per 1M tokens. That's 25% under Qwen3.8-Max, and Qwen's own testing puts average cost per task at $1.67, down from $4.43.
tokenrouter.com
2
243
Claude Fable 5.2 is now live on TokenRouter.
It's priced at $8 input and $40 output per 1M tokens, 20% below Fable 5.1, with cache reads at $0.20.
At High effort it scores 45.3% on FrontierCode 2.0. Opus 5.5 needs Max effort to reach 45.5%, and Fable gets there for about half the cost per task.
tokenrouter.com/models
1
1
414
The 🐰 is out of the bag. It’s MiniMax M3.1 Flash from @MiniMax_AI, and it’s now on TokenRouter.
M3.1 Flash scores 73.7% on DeepSWE v1.1, so it can burrow into a real codebase and come back with working fixes. On AutomationBench-AA it reaches 62.2, a test of business workflows that jump across several SaaS tools before the job is done. Output costs $0.60 per million tokens, which is not many carrots for a model this fast on its feet.
Let it loose on TokenRouter: tokenrouter.com/models
Claude Haiku 5.5 (@AnthropicAI) is now live on TokenRouter.
2
3
1,329
Get ready for SF @Techweek_ with TokenRouter! We're hosting two events in Palo Alto, and we'd love to see you there.
On Monday, Oct 5, we're kicking off the week with @inco_ai at the Token Happy Hour. Come talk inference speed with the team currently ranked #1 in output speed on Artificial Analysis for leading open-weight models.
On Saturday, Oct 10, we're co-hosting Decision-Grade AI at Stanford with Ezio. The day opens with Kiana Jafari of the Stanford Center for AI Safety, followed by a panel with leaders from @OpenAI, @Microsoft, @IBM and @Stanford, and then a buildathon focused on making LLM outputs trustworthy enough to act on.
Spots are limited. Save yours now:
>Token Happy Hour: partiful.com/e/JYHIeSugbKBQh…
>Decision-Grade AI @ Stanford: partiful.com/e/VVaPHqxrY0UYM…
See you next week.
Claude Sonnet 5.5 (@AnthropicAI) is now available on TokenRouter.
6
941
Running GLM-5.3? You're paying too much.
> 60% off on TokenRouter at "z-ai/glm-5.3".
tokenrouter.com
GPT-6 Sol and GPT-6 Luna (@OpenAI) are now available on TokenRouter.
GPT-6 Sol is priced at $2 per 1M input tokens and $10 per 1M output tokens, less than half the cost of GPT-5.5. GPT-6 Luna starts at $0.10 input and $0.50 output.
Both are accessible through a single API alongside 140+ models, with automatic failover and zero data retention.
Learn more at tokenrouter.com
666
Inco AI is entering public beta with the ⚡️ fastest ⚡️ #inference endpoints for #Kimi K3, #MiniMax M3, #GLM 5.3, and #GLM 5.3 Flash.
Ranking Number 1 on the respective @ArtificialAnlys provider leaderboard for output speed.
Read more: inco.ai/blog/inco-platform-a…
Gemini 3.8 Flash (@GeminiApp) is now available on TokenRouter.
3
15
1,880
GLM 5.3. Day 0. Running on our own GB300s right now.
TokenRouter now serves GLM 5.3, available from day 0 on our own GB300 cluster. Developed with technology partner @inco_ai, whose DFlash2 drafter and NVFP4 Blackwell checkpoint are live in the serving path.
Congratulations to @Zai_org on today's release.
Our blog post: tokenrouter.com/blog/tokenro…
Inco AI's technical results, DFlash2 and NVFP4:
c58eb90c-inco-ai.inco-ai.wor…
Weights(huggingface): huggingface.co/zai-org/GLM-5…
Try GLM 5.3 on TokenRouter: tokenrouter.com/models/
👀 Free for every account through Sep 4.
Wan 3.0 is now live on TokenRouter! The latest video generation engine from @Alibaba_Qwen with stronger subject consistency, and flexible text, image, and reference controls.
Learn more: tokenrouter.com/models
2
6
1,474
GLM 5.3 flash is now live on TokenRouter!
Learn more: tokenrouter.com/models






















