OCT 20, 2026 · MAPLEHILL LABS · WEEKLY · MARKET
Model market week of October 19, 2026
The market this week
New in the index (14):
- Mercury 3 (Inception) — from $0.100/MTok input
- DeepSeek V4.1 Flash 1017 (DeepSeek) — from $0.080/MTok input
- Pareto 26.10 (Unbiased) — from $0.800/MTok input
- Glimpse (Elorian) — from $0.600/MTok input
- Beam 1.0 501B A23B (Reflection) — from $0.250/MTok input
- Moondream 3 Preview (Moondream) — from $0.300/MTok input
- Moondream 3.1 (Moondream) — from $0.300/MTok input
- Muse Spark Open (Meta) — from $1.25/MTok input
- Nemotron 3.5 Super VL (NVIDIA) — from $0.100/MTok input
- Pareto 26.9 (Unbiased) — from $2.50/MTok input
- LFM2.5-24B-A2B (Liquid) — from $0.050/MTok input
- Gemini 4 Argon (Google) — from $2.00/MTok input
- Summarize (Scaledown) — from $0.040/MTok input
- Extract (Scaledown) — from $0.040/MTok input
Price moves — largest 8 of 30 observed moves:
- DeepSeek V4 Flash 0423 output: $0.024 → $0.136/MTok (+466.7%)
- DeepSeek V4 Pro 0813 output: $0.400 → $1.98/MTok (+395.0%)
- GLM 5.3 Flash input: $0.013 → $0.060/MTok (+380.0%)
- DeepSeek V4 Pro 0423 input: $0.177 → $0.450/MTok (+153.5%)
- GPT-5.6 Sol Pro input: $2.00 → $4.00/MTok (+100.0%)
- GPT-5.6 Sol Pro output: $10.00 → $20.00/MTok (+100.0%)
- GLM 5.2 output: $1.80 → $0.594/MTok (-67.0%)
- Solar Pro 4 input: $0.090 → $0.150/MTok (+66.7%)
Provider availability:
- DeepSeek V4.1 Flash 1017: now on Alibaba, AtlasCloud, Decart, GMICloud, Crusoe, Morph, Together, Wafer
- MiMo-V2.6-Flash: now on Relace
- DeepSeek V4.1 Flash: now on Relace, Nebius, PrimeIntellect; left Relace, Cloudflare, DeepSeek
- Qwen3.8 27B: now on Modal, Tenstorrent; left Tenstorrent
- GLM 5.2: now on Relace, Relace; left Relace, InferenceNet
- GLM 4.7 Flash: left Venice
- R1 0528: left Novita
- Llama 4 Maverick: left Novita
Most-used models (OpenRouter tokens, trailing 7 days):
- DeepSeek V4.1 Flash — 29.6T tokens
- GLM 5.3 Flash — 11.9T tokens
- MiMo-V2.6-Flash — 9.6T tokens
- Step 5 Preview — 8.2T tokens
- Hy4 preview — 7.8T tokens
Measured note: z-ai/glm-5.3-flash on pareto-inference slowed week-over-week — median TTFT 1.27s → 3.35s (≥3 daily probes each week).
Around the industry
- Google released Gemini 4 Argon globally on October 13, 2026, through its API, AI Studio, Gemini platform, Gemini app, and Search AI Mode. [1]
- Google priced Gemini 4 Argon at an introductory $2 per million input tokens and $10 per million output tokens through November 30, 2026. [1]
- Google published the Gemini 4 Argon model card on October 13, 2026. [3]
- OpenAI retired GPT-5.5 from ChatGPT, ChatGPT Work, and Codex on October 14, 2026, while retaining API access. [6]
- OpenAI began rolling out GPT-Live-1.1 and GPT-Live-1.1 Quip in ChatGPT Voice on October 15, 2026. [8]
- OpenAI introduced GPT-6 with Intelligent UI in ChatGPT, beginning rollout to paid plans on October 7 and expanding to Free and Go users on October 8, 2026. [10]
- OpenAI published an October 19, 2026 update describing improvements to ChatGPT coding support for low-resource programming languages, including SQLite and COBOL migration. [7]
Assembled automatically from ModelIndex's observed data and cited external sources. Sourced claims link to their citations; our own figures link to the live pages they come from.