As of August 20, 2026, Hy3 is 76% cheaper per million tokens than GLM-4.6. GLM-4.6 is $0.60 / $2.20 per million input/output tokens versus Hy3 at $0.13 / $0.53. Sources: https://docs.z.ai/guides/overview/pricing and https://www.tencentcloud.com/document/product/1300/78937
Sources: Z AI pricing (https://docs.z.ai/guides/overview/pricing); Tencent pricing (https://www.tencentcloud.com/document/product/1300/78937)
GLM-4.6
Z AI
Input
$0.60
Output
$2.20
Hy3
Tencent
Input
$0.13
Output
$0.53
Token pricing, specs, and benchmarks side by side
GLM-4.6Hy3Verdict
Hy3 offers both lower pricing and stronger benchmark performance across the board, making it the clear value leader in this comparison.
As of August 20, 2026, Hy3 is 76% cheaper on blended LLM API pricing. Hy3 costs $0.13 per million input tokens and $0.53 per million output tokens ($0.66 blended 1M-in + 1M-out). GLM-4.6 costs $0.60 / $2.20 per million tokens ($2.80 blended). Sources: https://docs.z.ai/guides/overview/pricing and https://www.tencentcloud.com/document/product/1300/78937. The cheapest LLM for your app still depends on how many output tokens you generate.
GLM-4.6 Z AI API pricing is $0.60 input and $2.20 output per million tokens via Z AI. Hy3 Tencent API pricing is $0.13 input and $0.53 output per million tokens via Tencent. Use those four numbers, not ChatGPT Plus or Claude Pro subscription prices, when you are comparing APIs.
Hy3 leads on SWE-bench Verified: GLM-4.6 at 68% vs Hy3 at 78%. SWE-bench measures real GitHub fixes; LiveCodeBench measures contest programming.
GLM-4.6 supports a 203K token context window, while Hy3 supports 262K tokens. Hy3 offers a larger context window. Filling a larger window bills more input tokens, so the cheaper-per-million model can still cost more on long documents.
Hy3 leads on 5 of 5 shared benchmarks versus GLM-4.6's 0 wins. Check the comparison table for GPQA Diamond, SWE-bench, MMLU, HLE, and the other evals we track. Token price and benchmark score together are the usual LLM comparison, not either number alone.
Both GLM-4.6 and Hy3 are production API models. For cost-sensitive production traffic, Hy3 has better economics on the blended 1M-in + 1M-out scale. For maximum capability, weight the benchmark table for your domain. Many production stacks route cheap models for drafts and a frontier model for the hard turn, which is usually cheaper than sending everything to the expensive API.
Yes. Call GLM-4.6 through Z AI and Hy3 through Tencent with separate API keys, or through a gateway that already wraps both. Keep prompts in tokens, not characters, when you estimate the invoice. AnotherWrapper templates can swap providers without rewriting auth and billing.
List rates are the latest published Z AI API pricing and Tencent API pricing, quoted per million tokens. Official rate sheets: https://docs.z.ai/guides/overview/pricing and https://www.tencentcloud.com/document/product/1300/78937. Prompt caching, batch APIs, and committed-use discounts can change the invoice. Benchmark scores are from official publications and independent evals, updated as new numbers land.
GLM-4.6 supports up to 131K output tokens per request, while Hy3 supports up to an unspecified number of output tokens. Output tokens are usually the expensive half of API pricing, so a higher max output cap is a capability, not a discount.
GLM-4.6 has a measured throughput of approximately 85 tokens/second. Throughput data for Hy3 is not yet available. Throughput (tokens per second) and time-to-first-token are why two models with the same per-million-token price can still feel different in chat and agents.
Use the model pickers at the top of this LLM comparison. Search any model in the index and the URL updates. Open GLM-4.6 or Hy3 from the sidebar for the single-model API pricing page, or go back to the LLM pricing table to sort by cheapest blended cost.
This comparison covers API pricing, context window, throughput, and benchmark scores. GLM-4.6 costs $0.60/M input and $2.20/M output. Hy3 costs $0.13/M input and $0.53/M output.
Rank both on the LLM leaderboard, SWE-bench, and Humanity's Last Exam.
The index
280 models across 33 providers. Search or jump to a lab — every model page stays linked here.
Next step
Auth, billing, and the API layer are already decided. Fork 8 finished AI apps, or have us build the first version with you.
