Tokenomics

We approximate task-completion effectiveness by weighting intelligence: every 5 points doubles the quality weight—not a measured success rate.

xychart-beta
    title "Intelligence → exponential quality weight"
    x-axis "Intelligence" [0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60]
    y-axis "Quality weight (50 = 1)" 0 --> 4
    line [0.0009765625, 0.001953125, 0.00390625, 0.0078125, 0.015625, 0.03125, 0.0625, 0.125, 0.25, 0.5, 1, 2, 4]

How the scores work#

• Provider / plan: who sells it and which plan; PAYG is pay-as-you-go API.
• Monthly $: plan price per month in USD before tax, monthly billing; annual variants and older measurement windows are left out.
• Model / effort: the benchmarked model and its reasoning-effort setting.
• Intelligence: Slopalytics benchmark score.
• Exponential quality weight: 2^((intelligence − 50)/5), so 50 = 1 and every 5 points doubles it; linear tables use intelligence itself.
• API $/task (C): Slopalytics API cost to complete one benchmark task.
• API-equivalent/$ (R): API dollars of usage each plan dollar buys per month (weekly quota × 52/12 plus API credits, ÷ price); PAYG = 1. Credit plans assume full use at 5,000 fresh + 95,000 cached input + 2,000 output tokens per task.
• Raw tasks/$: R/C, tasks per dollar ignoring quality.
• Task score ↑: quality weight × R/C, quality-weighted tasks per plan dollar including the plan discount; Tasks / $ tables sort by this.
• Intelligence/$ ↑: I × quality weight × R/C, the intelligence delivered across the tasks a plan dollar buys; Intelligence / $ tables sort by this.
• Plans by API-equivalent $: every plan with the API dollars of usage it buys per month (R × price), sorted by the median Exponential task score; both medians cover only the plan's Pareto models (no other model on the plan is at least as smart and no dearer per task), listed under the table; pay-as-you-go API is left out.
• Models by API $: one row per model at official API list prices (R = 1), keeping the reasoning effort with the best Exponential intelligence/$ and sorted by it; plans and OpenRouter routes are left out.
• Cheap vs smart model: per provider, the cheapest-per-task model against the smartest, priced on the named plan; Cost × (smart vs cheap) is the smart model's cost per task ÷ the cheap model's, and Intelligence × (smart vs cheap) is its intelligence ÷ the cheap model's.
• Notable: three generated one-liners under each table: the leader's margin, where the smartest row lands and why, and the best row from a different provider.
• Pareto: ★ when no row in the same pool is at least as smart and at least as cheap per task (C/R).
• Evidence: how the row was estimated (benchmark, modeled credits or modeled transfer), linked to its source note; ND means no data.

Best Exponential Tasks / $#

Notable

• #1 Claude Haiku 5.5 (medium) (Anthropic / Max 20x) leads #2 Claude Haiku 5.5 (medium) (Anthropic / Pro) by 1.06×.
• Smartest row Claude Opus 5.5 (high with fallback) (intelligence 53.58) ranks #8: F 14.2× · R 1× ÷ C 38.8× vs #1.
• Best non-Anthropic row: GPT-6 Luna (low) (OpenAI / ChatGPT Plus — FinOps measurement) at #4, 1.55× behind #1.

Best Exponential Intelligence / $#

Notable

• #1 Claude Haiku 5.5 (xhigh) (Anthropic / Max 20x) leads #2 Claude Haiku 5.5 (xhigh) (Anthropic / Pro) by 1.06×.
• Smartest row Claude Opus 5.5 (high with fallback) (intelligence 53.58) ranks #7: I 1.3× · F 5.53× · R 1× ÷ C 14.7× vs #1.
• Best non-Anthropic row: GPT-6.1 Sol (medium) (OpenAI / ChatGPT Plus — FinOps measurement) at #4, 1.62× behind #1.

Best Open Exponential Tasks / $#

Notable

• #1 MiMo-V2.6-Pro (Xiaomi / Token Plan Max monthly) leads #2 MiMo-V2.6-Flash (Xiaomi / Token Plan Max monthly) by 1.56×.
• The smartest row, MiMo-V2.6-Pro (intelligence 46.32), is also #1.
• Best non-Xiaomi row: GLM-5.3-Flash (PAYG API) at #3, 4.73× behind #1.

Best Open Exponential Intelligence / $#

Notable

• #1 MiMo-V2.6-Pro (Xiaomi / Token Plan Max monthly) leads #2 MiMo-V2.6-Flash (Xiaomi / Token Plan Max monthly) by 1.91×.
• The smartest row, MiMo-V2.6-Pro (intelligence 46.32), is also #1.
• Best non-Xiaomi row: GLM-5.3-Flash (PAYG API) at #3, 5.24× behind #1.

Best Linear Tasks / $#

Notable

• #1 GPT-6 Luna (low) (OpenAI / ChatGPT Plus — FinOps measurement) leads #2 GPT-6 Luna (low) (OpenAI / Pro $100 — Reset Oct 3) by 1.01×.
• Smartest row Claude Fable 5.1 (low with fallback) (intelligence 46.82) ranks #12: F 2.17× · R 2.68× ÷ C 525× vs #1.
• Best non-OpenAI row: Claude Haiku 5.5 (low) (Anthropic / Max 20x) at #4, 1.48× behind #1.

Best Linear Intelligence / $#

Notable

• #1 GPT-6 Luna (low) (OpenAI / ChatGPT Plus — FinOps measurement) leads #2 GPT-6 Luna (low) (OpenAI / Pro $100 — Reset Oct 3) by 1.01×.
• Smartest row Claude Fable 5.1 (low with fallback) (intelligence 46.82) ranks #12: I 2.17× · F 2.17× · R 2.68× ÷ C 525× vs #1.
• Best non-OpenAI row: Claude Haiku 5.5 (low) (Anthropic / Max 20x) at #4, 1.08× behind #1.

Best Open Linear Tasks / $#

Notable

• #1 gpt-oss-20b (high) (PAYG API) leads #2 MiMo-V2.6-Flash (Xiaomi / Token Plan Max monthly) by 1.03×.
• Smartest row MiMo-V2.6-Pro (intelligence 46.32) ranks #3: F 5.17× · R 1.19× ÷ C 10.7× vs #1.
• Best non-OpenAI row: MiMo-V2.6-Flash (Xiaomi / Token Plan Max monthly) at #2, 1.03× behind #1.

Best Open Linear Intelligence / $#

Notable

• #1 MiMo-V2.6-Flash (Xiaomi / Token Plan Max monthly) leads #2 MiMo-V2.6-Pro (Xiaomi / Token Plan Max monthly) by 1.38×.
• Smartest row MiMo-V2.6-Pro (intelligence 46.32) ranks #2: I 1.22× · F 1.22× · R 1.04× ÷ C 2.14× vs #1.
• Best non-Xiaomi row: gpt-oss-20b (high) (PAYG API) at #3, 4.09× behind #1.

Plans by API-equivalent $#

Monthly billing, latest measurement only; sorted by median Exponential Tasks / $. Medians are over the plan's Pareto models, listed below the table.

Models used in each median (cheapest per task first):

• OpenAI (Plus, Pro 100, Pro 200): GPT-6 Luna (low), GPT-6 Luna (medium), GPT-6 Luna (high), GPT-6 Luna (xhigh), GPT-6 Luna (max), GPT-6.1 Sol (low), GPT-6.1 Sol (medium), GPT-6.1 Sol (high), GPT-6.1 Sol (xhigh), GPT-6.1 Sol (max), GPT-6 Astra (xhigh), GPT-6 Astra (max)
• Anthropic (Max 20x, Pro, Team Standard, Team Premium): Claude Haiku 5.5 (low), Claude Haiku 5.5 (medium), Claude Haiku 5.5 (high), Claude Haiku 5.5 (xhigh), Claude Haiku 5.5 (max), Claude Sonnet 5.5 (high with fallback), Claude Opus 5.5 (medium with fallback), Claude Opus 5.5 (high with fallback), Claude Opus 5.5 (xhigh with fallback), Claude Sonnet 5.5 (max with fallback), Claude Opus 5.5 (max with fallback)
• Xiaomi (Token Plan Max monthly, Token Plan Pro monthly, Token Plan Standard monthly, Token Plan Lite monthly): MiMo-V2.6-Flash, MiMo-V2.6-Pro

Models by API $#

Official API list prices only (no plans, no OpenRouter); one row per model, at its best reasoning effort by Exponential intelligence / $.

Cheap vs smart model#

Evidence details#

1. Anthropic / Claude Haiku 5.5 (medium) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
2. Anthropic / Claude Haiku 5.5 (medium) / Pro: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=736.67/month; B=0; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
3. Anthropic / Claude Haiku 5.5 (medium) / Team Standard: claude, claude_credits, wokey; 2026 10/01 – 10/08; measured mixed-workload A=749.67/month; B=20; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
4. OpenAI / GPT-6 Luna (low) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
5. OpenAI / GPT-6 Luna (low) / Pro $100 — Reset Oct 3: openai, inventive; 2026-10-04; measured $3.31/1% weekly quota; A=1,434.33/month; current-model repricing/allowance identity unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
6. OpenAI / GPT-6 Luna (low) / Pro $200 — Reset Oct 9 (Oct 2–4 so far): openai, inventive; 2026-10-04; measured $6.41/1% weekly quota; A=2,777.67/month; current-model repricing/allowance identity unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
7. OpenAI / GPT-6.1 Sol (medium) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
8. Anthropic / Claude Opus 5.5 (high with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
9. Anthropic / Claude Sonnet 5.5 (high with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
10. Anthropic / Claude Fable 5.1 (high with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
11. OpenAI / GPT-6 Astra (low) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
12. Xiaomi / MiMo-V2.6-Pro / Token Plan Max monthly: mimo, mimo_api; modeled fixed workload; deductions fresh/cache/output=300 / 2.5 / 600 credits/token; A=118.83/month; benchmark-throughput transfer is hypothetical
13. OpenAI / GPT-5.6 Terra (medium) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
14. Xiaomi / MiMo-V2.6-Flash / Token Plan Max monthly: mimo, mimo_api; modeled fixed workload; deductions fresh/cache/output=100 / 2 / 200 credits/token; A=114.8/month; benchmark-throughput transfer is hypothetical
15. Z AI / GLM-5.3-Flash / Benchmark API rate card; official headline rates match: benchmark, aa; source benchmark measurement; headline input/output verified, cache/context/fallback may differ; benchmark input/cache/output=0.15 / ND / 0.5
16. Google / Gemini 4 Argon (high) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=2 / ND / 10
17. Meta / Muse Spark 1.3 (max) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=1.25 / ND / 4.25
18. Meta / Muse Wonder (max) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=4 / ND / 20
19. SpaceXAI / Grok 4.7 (low) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=2 / ND / 6
20. Anthropic / Claude Haiku 5.5 (xhigh) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
21. Anthropic / Claude Haiku 5.5 (xhigh) / Pro: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=736.67/month; B=0; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
22. Anthropic / Claude Haiku 5.5 (xhigh) / Team Standard: claude, claude_credits, wokey; 2026 10/01 – 10/08; measured mixed-workload A=749.67/month; B=20; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
23. OpenAI / GPT-6.1 Sol (medium) / Pro $100 — Reset Oct 3: openai, inventive; 2026-10-04; measured $3.31/1% weekly quota; A=1,434.33/month; current-model repricing/allowance identity unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
24. OpenAI / GPT-6.1 Sol (medium) / Pro $200 — Reset Oct 9 (Oct 2–4 so far): openai, inventive; 2026-10-04; measured $6.41/1% weekly quota; A=2,777.67/month; current-model repricing/allowance identity unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
25. OpenAI / GPT-6 Luna (max) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
26. OpenAI / GPT-6 Astra (high) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
27. OpenAI / GPT-5.6 Terra (high) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
28. OpenAI / gpt-oss-20b (high) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=0.07 / ND / 0.18
29. Z AI / GLM-5.3 (max) / Benchmark API rate card; official headline rates match: benchmark, aa; source benchmark measurement; headline input/output verified, cache/context/fallback may differ; benchmark input/cache/output=1.4 / ND / 4.4
30. Kimi / Kimi K3 (max) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=3 / ND / 15
31. Kimi / Kimi K2.7 Code / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=0.95 / ND / 4
32. OpenAI / gpt-oss-120b (high) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=0.15 / ND / 0.59
33. Anthropic / Claude Haiku 5.5 (low) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
34. Anthropic / Claude Haiku 5.5 (low) / Pro: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=736.67/month; B=0; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
35. Anthropic / Claude Haiku 5.5 (low) / Team Standard: claude, claude_credits, wokey; 2026 10/01 – 10/08; measured mixed-workload A=749.67/month; B=20; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
36. OpenAI / GPT-6.1 Sol (low) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
37. Anthropic / Claude Sonnet 5.5 (low with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
38. Anthropic / Claude Opus 5.5 (low with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
39. OpenAI / GPT-5.6 Terra (low) / ChatGPT Plus — FinOps measurement: openai, finops; October 10, 2026; measured A=290.33/month; workload/rate-card transfer to current benchmark efforts unverified; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
40. Anthropic / Claude Fable 5.1 (low with fallback) / Max 20x: claude, claude_credits, wokey; 2026 10/05 – 10/12; measured mixed-workload A=7,587.67/month; B=200; R is workload-specific, NOT a per-model benchmark quota; modeled fixed-R transfer across models/efforts; named on official provider page; tier-specific access and model-specific quota consumption unverified
41. Google / Gemini 3.5 Flash-Lite / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=0.3 / ND / 2.5
42. Google / Gemini 3.8 Flash (medium) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=0.75 / ND / 3.75
43. Meta / Muse Spark 1.3 (xhigh) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=1.25 / ND / 4.25
44. Z AI / GLM-5.3 (low) / Benchmark API rate card; official headline rates match: benchmark, aa; source benchmark measurement; headline input/output verified, cache/context/fallback may differ; benchmark input/cache/output=1.4 / ND / 4.4
45. Kimi / Kimi K3 (low) / Benchmark API rate card: benchmark, aa; source benchmark measurement; published benchmark cost; current provider billing not independently matched; benchmark input/cache/output=3 / ND / 15

Source health and dates#

• [benchmark] https://slopalytics.com/ — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [aa] https://artificialanalysis.ai/models — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [router] https://openrouter.ai/api/v1/models — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [openai] https://learn.chatgpt.com/docs/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [openai_tiers] https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers — FETCH FAILED: HTTPError: HTTP Error 403: Forbidden; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [finops] https://finopsllm.com/research/codex-usage-limits-chatgpt-pro — fetched successfully; publication/update: October 10, 2026; fetched 2026-10-18T01:13:38+00:00.
• [inventive] https://inventivehq.com/blog/codex-subscription-usage-measured — fetched successfully; publication/update: 2026-10-04; fetched 2026-10-18T01:13:38+00:00.
• [claude] https://claude.com/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [claude_credits] https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [claude_api] https://platform.claude.com/docs/en/about-claude/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:38+00:00.
• [wokey] https://wokey.ai/claude-usage-limits — fetched successfully; publication/update: 2026-10-12; fetched 2026-10-18T01:13:39+00:00.
• [google] https://one.google.com/about/google-ai-plans/ — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:39+00:00.
• [google_plans] https://gemini.google/subscriptions/ — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:39+00:00.
• [google_api] https://ai.google.dev/gemini-api/docs/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:39+00:00.
• [x] https://x.ai/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:40+00:00.
• [x_news] https://x.ai/news/x-subscription — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:40+00:00.
• [cursor] https://cursor.com/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:40+00:00.
• [muse] https://dev.meta.ai/products/muse-code — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:40+00:00.
• [kimi] https://www.kimi.ai/help/membership/membership-pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:40+00:00.
• [kimi_limits] https://www.kimi.ai/help/membership/membership-overview — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:41+00:00.
• [kimi_api] https://platform.kimi.ai/docs/pricing/chat — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:41+00:00.
• [mimo] https://mimo.mi.com/docs/en-US/price/token-plan — fetched successfully; publication/update: September 21, 2026; fetched 2026-10-18T01:13:41+00:00.
• [mimo_api] https://mimo.mi.com/docs/en-US/price/pay-as-you-go — fetched successfully; publication/update: October 08, 2026; fetched 2026-10-18T01:13:41+00:00.
• [glm] https://docs.z.ai/devpack/overview — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:42+00:00.
• [glm_plans] https://z.ai/subscribe — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:43+00:00.
• [glm_faq] https://docs.z.ai/devpack/faq — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:43+00:00.
• [glm_api] https://docs.z.ai/guides/overview/pricing — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:13:43+00:00.
• [benchmark_script_0] https://slopalytics.com/assets/index-C2DcVdUk.js — fetched successfully; publication/update: publication date not exposed; fetched 2026-10-18T01:14:13+00:00.

Parsing/coverage warnings#

• GLM: live subscription prices were not extractable from official HTML. Quotas and modeled API capacity are shown, but price-dependent scores are ND.
• openai_tiers: HTTPError: HTTP Error 403: Forbidden; dependent values are ND.