subvalue: AI coding plan changes

Gemini 4 Hydrogen released

Thu, 12 Nov 2026 12:00:00 GMT

Intelligence Index 53.9 · $1.5 / $7.5 per million tokens Google AI Pro: <1 → 293 turns of balanced work a month Google AI Ultra 5×: 1.9k → 2.8k turns of balanced work a month Google AI Ultra 20×: 7.4k → 11k turns of balanced work a month

Gemini 4 Carbon comes to Copilot Pro+ and Max

Tue, 10 Nov 2026 12:00:00 GMT

Copilot Pro+: 154 → 210 turns of balanced work a month Copilot Max: 441 → 601 turns of balanced work a month

New best model: Fable 5.2; Claude Fable 5.2 released; Muse Ember released

Mon, 09 Nov 2026 12:00:00 GMT

64.6 on the Intelligence Index. The score each kind of work needs moves up with it. Intelligence Index 64.6 · $8 / $40 per million tokens Intelligence Index 36 · $0.1 / $0.5 per million tokens Cursor Ultra: 12 → 2.4 turns of balanced work a month Kimi Ultra: 29 → 6.7 turns of balanced work a month Kimi Max: 15 → 3.3 turns of balanced work a month OpenCode Go: 37 → 8.9 turns of balanced work a month OpenCode Go Plus: 147 → 36 turns of balanced work a month

Mistral Large 4 released; ChatGPT Pro allowances doubled

Thu, 29 Oct 2026 12:00:00 GMT

Intelligence Index 26.1 · $0.7 / $2.1 per million tokens Pro 100 to 10×, Pro 200 back to 20×, Pro 500 to 50× Plus. Plus unchanged. ChatGPT Pro 100: 2.9k → 5.9k turns of balanced work a month ChatGPT Pro 200: 5.2k → 10k turns of balanced work a month ChatGPT Pro 500: 15k → 29k turns of balanced work a month

New best model: Gemini 4 Carbon; Gemini 4 Carbon released

Wed, 28 Oct 2026 12:00:00 GMT

60 on the Intelligence Index. The score each kind of work needs moves up with it. Intelligence Index 60 · $2 / $10 per million tokens Google AI Ultra 20×: 4.5k → 8.5k turns of balanced work a month Google AI Ultra 5×: 1.1k → 2.1k turns of balanced work a month Cursor Ultra: 16 → 12 turns of balanced work a month OpenCode Go: 50 → 37 turns of balanced work a month OpenCode Go Plus: 200 → 147 turns of balanced work a month

Kimi K3.1 Pro uses about ten times the quota per token

Tue, 27 Oct 2026 12:00:00 GMT

K3 is unchanged. Kimi Plus: 27 → 2.7 turns of balanced work a month Kimi Max: 189 → 19 turns of balanced work a month Kimi Ultra: 378 → 38 turns of balanced work a month Kimi Pro: 51 → 5.1 turns of balanced work a month

GPT-6.1 Terra released; Kimi Ultra launches; Kimi's new plans halve the allowance per tier

Mon, 26 Oct 2026 12:00:00 GMT

Intelligence Index 51.3 · $1 / $5 per million tokens Twice Max's agent credits. Moderato, Allegretto and Allegro become Plus, Pro and Max at the same prices with half the tokens, per Real API Pricing's reading. Kimi Ultra: no figure → 378 turns of balanced work a month Kimi Pro: 102 → 51 turns of balanced work a month Kimi Max: 378 → 189 turns of balanced work a month Kimi Plus: 54 → 27 turns of balanced work a month ChatGPT Pro 500: 12k → 17k turns of balanced work a month

MiniMax-M3.1-Flash released; Kimi K3.1 Pro released

Fri, 23 Oct 2026 12:00:00 GMT

Intelligence Index 29.1 · $0.15 / $0.6 per million tokens Intelligence Index 46.5 · $2 / $10 per million tokens Kimi Max: <1 → 378 turns of balanced work a month Kimi Pro: <1 → 102 turns of balanced work a month Kimi Plus: <1 → 54 turns of balanced work a month

Gemini 4 Argon arrives in Antigravity, for Ultra

Tue, 13 Oct 2026 12:00:00 GMT

Google AI Ultra 20×: 44 → 4.5k turns of balanced work a month Google AI Ultra 5×: 11 → 1.1k turns of balanced work a month

Muse Wonder released; Muse Wonder Full released; Cursor's new plans replace Pro, Pro+ and the old Ultra

Mon, 12 Oct 2026 12:00:00 GMT

Intelligence Index 43.6 · $4 / $20 per million tokens Intelligence Index 50.1 · $4 / $20 per million tokens Plus $8 (Grok only), Premium $30, Super $100, Ultra $200, with weekly pools. The plans they replaced are not shown. Cursor Ultra: no figure → 16 turns of balanced work a month Muse Code Maximum: 37 → 1.2k turns of balanced work a month Muse Code Power: 9.4 → 308 turns of balanced work a month Muse Code High: 2.3 → 12 turns of balanced work a month

Mistral Large 4 Preview released; Claude Haiku 5.5 released

Wed, 07 Oct 2026 12:00:00 GMT

Intelligence Index 25.3 · $0.7 / $2.1 per million tokens Intelligence Index 36.5 · $0.1 / $0.5 per million tokens Devin Max: 4.5k → 9.7k turns of balanced work a month Devin Pro: 264 → 571 turns of balanced work a month Claude Max: 8.8k → 18k turns of balanced work a month Claude Pro: 1.6k → 3.3k turns of balanced work a month Claude Ultra: 17k → 34k turns of balanced work a month

Gemini 4 Argon released

Wed, 30 Sep 2026 12:00:00 GMT

Intelligence Index 50.4 · $2 / $10 per million tokens OpenCode Go Plus: 2 → 157 turns of balanced work a month OpenCode Go: 1 → 39 turns of balanced work a month

GPT-6.1 Sol released; ChatGPT Pro 500 launches; ChatGPT Pro 200 cut from 20× to 10× Plus

Tue, 29 Sep 2026 12:00:00 GMT

Intelligence Index 54.6 · $2 / $10 per million tokens 25× Plus usage at launch, with Ultrafast mode. Reported from subscriber emails; OpenAI did not publish it. ChatGPT Pro 500: no figure → 12k turns of balanced work a month ChatGPT Pro 100: 790 → 2.3k turns of balanced work a month ChatGPT Plus: 159 → 470 turns of balanced work a month ChatGPT Pro 200: 2.8k → 4.2k turns of balanced work a month Copilot Max: 382 → 463 turns of balanced work a month

Claude Sonnet 5.5 released; OpenCode Go Plus launches at $40

Mon, 28 Sep 2026 12:00:00 GMT

Intelligence Index 54.5 · $2 / $10 per million tokens Per-model allowances two to eight times Go's. Claude Max: 5.5k → 8.8k turns of balanced work a month Claude Ultra: 11k → 17k turns of balanced work a month Claude Pro: 1k → 1.6k turns of balanced work a month Devin Max: 2.8k → 4.1k turns of balanced work a month Devin Pro: 164 → 240 turns of balanced work a month

Grok 4.7 released; New best model: Opus 5.5; Claude Opus 5.5 released; GPT-6 Luna released; GPT-6 Sol released

Tue, 22 Sep 2026 12:00:00 GMT

Intelligence Index 35.5 · $2 / $6 per million tokens 59.1 on the Intelligence Index. The score each kind of work needs moves up with it. Intelligence Index 59.1 · $4 / $20 per million tokens Intelligence Index 27.9 · $0.1 / $0.5 per million tokens Intelligence Index 44.2 · $2 / $10 per million tokens Claude Pro: 111 → 1k turns of balanced work a month Claude Max: 1.4k → 5.5k turns of balanced work a month Claude Ultra: 2.9k → 11k turns of balanced work a month ChatGPT Plus: 73 → 159 turns of balanced work a month ChatGPT Pro 100: 366 → 790 turns of balanced work a month

New best model: GPT-6 Astra; GPT-6 Astra released

Thu, 03 Sep 2026 12:00:00 GMT

56.9 on the Intelligence Index. The score each kind of work needs moves up with it. Intelligence Index 56.9 · $10 / $50 per million tokens GLM Coding Plan Pro: 29 → 2.7 turns of balanced work a month Google AI Pro: 51 → 4.7 turns of balanced work a month Google AI Ultra 20×: 1k → 93 turns of balanced work a month GLM Coding Plan Max: 68 → 6.3 turns of balanced work a month Google AI Ultra 5×: 252 → 23 turns of balanced work a month

Gemini 3.8 Flash released; Qwen3.8 Max (0902) released; Muse Spark 1.3 released

Wed, 02 Sep 2026 12:00:00 GMT

Intelligence Index 30.8 · $0.75 / $3.75 per million tokens Intelligence Index 33.7 · $2 / $6 per million tokens Intelligence Index 36.6 · $1.25 / $4.25 per million tokens Google AI Pro: no figure → 51 turns of balanced work a month Google AI Ultra 5×: no figure → 252 turns of balanced work a month Google AI Ultra 20×: no figure → 1k turns of balanced work a month Muse Code High: no figure → 48 turns of balanced work a month Muse Code Power: no figure → 190 turns of balanced work a month

New best model: Fable 5.1; Claude Fable 5.1 released

Tue, 01 Sep 2026 12:00:00 GMT

50 on the Intelligence Index. The score each kind of work needs moves up with it. Intelligence Index 50 · $10 / $50 per million tokens ChatGPT Plus: 96 → 13 turns of balanced work a month ChatGPT Pro 100: 478 → 64 turns of balanced work a month ChatGPT Pro 200: 1.7k → 228 turns of balanced work a month OpenCode Go: 44 → 6.1 turns of balanced work a month GLM Coding Plan Max: 477 → 68 turns of balanced work a month

GLM-5.3-Flash released

Wed, 26 Aug 2026 12:00:00 GMT

Intelligence Index 27.8 · $0.15 / $0.5 per million tokens OpenCode Go: 5 → 44 turns of balanced work a month GLM Coding Plan Pro: 179 → 205 turns of balanced work a month GLM Coding Plan Max: 418 → 477 turns of balanced work a month GLM Coding Plan Lite: 30 → 34 turns of balanced work a month

DeepSeek V4 Pro 0813 released; GLM-5.3 released

Tue, 18 Aug 2026 12:00:00 GMT

Intelligence Index 23.1 · $1.32 / $3.96 per million tokens Intelligence Index 31.7 · $1.4 / $4.4 per million tokens GLM Coding Plan Lite: no figure → 30 turns of balanced work a month GLM Coding Plan Pro: no figure → 179 turns of balanced work a month GLM Coding Plan Max: no figure → 418 turns of balanced work a month