AI News Feed
Market watch
Chicxulub Impact Crater May Have Been a Hotbed of Life for 8 Million YearsGoogle Opens Gemini 4 Argon to Vetted Cybersecurity Defenders FirstElder fraud grows as exposed data fuels SIM-swap and deepfake scams, 9to5Mac reportsDestro AI raises $8M to coordinate mixed robot fleets in enterprise warehousesHPE Tells Investors Networking Is Now Its Growth Engine After Juniper DealFujifilm opens $170 pre-orders for Instax Pal 2, a palm-size camera built around dials and filters
Chicxulub Impact Crater May Have Been a Hotbed of Life for 8 Million YearsGoogle Opens Gemini 4 Argon to Vetted Cybersecurity Defenders FirstElder fraud grows as exposed data fuels SIM-swap and deepfake scams, 9to5Mac reportsDestro AI raises $8M to coordinate mixed robot fleets in enterprise warehousesHPE Tells Investors Networking Is Now Its Growth Engine After Juniper DealFujifilm opens $170 pre-orders for Instax Pal 2, a palm-size camera built around dials and filters
Large Language Models

OpenAI Releases GPT-6.1 Sol, a Mid-Tier Model Priced at One-Fifth of Astra's Rates

OpenAI released GPT-6.1 Sol at one-fifth of Astra's rates, citing near-Astra coding and computer-use benchmarks.

AI News Feed · 2026-10-01 08:08

Pricing is $2 per million input tokens, $10 per million output tokens and $0.10 per million cached input tokens. OpenAI describes the cached rate as 50 percent below GPT-6 Sol's, and says cached reads now cost 5 percent of the uncached input rate, down from 10 percent on GPT-6 Sol. Cached pricing carries unusual weight for agent deployments, because agents resend the same system prompt, tool schemas and conversation history at every step.

The company now ships three GPT-6 tiers. GPT-6 Astra costs $10 for input, $50 for output and $1 for cached input per million tokens. GPT-6.1 Sol costs $2, $10 and $0.10. GPT-6 Luna costs $0.10, $0.50 and $0.01.

All performance figures below are vendor-reported in OpenAI's launch post; OpenAI states that competitor numbers came from public reports. On DeepSWE v1.1, a coding benchmark, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth of the cost and beats GPT-6 Sol's best score by 6.4 percentage points at a lower reasoning effort. On GDP.pdf, which tests answers over complex professional PDFs, Sol outperforms Claude Opus 5.5 with fallbacks at less than half the cost per task. On AutomationBench 1.0.6, Sol scores 2.2 points above Opus 5.5 at medium effort, at roughly one-third the cost.

On the OSWorld 2.0 offline set for computer use, Sol gains 7 points over GPT-6 Sol at maximum effort and lands within 2.1 points of Astra at roughly one-seventh the cost per task. On Terminal-Bench Science 0.1, Sol more than doubles GPT-6 Sol's score at maximum effort, at an average cost of $5.47 per task against $23.21 for Opus 5.5 and $23.80 for Astra. Astra still leads on that evaluation at 68.1 percent, and OpenAI recommends it for the hardest research work.

On factuality, OpenAI reports that at low reasoning effort the share of responses containing a factual error falls from 11.4 percent to 7.7 percent, a reduction of about 32 percent. The evaluation uses deliberately difficult conversations in which users had flagged earlier model errors.

The model page lists a context window of 1,050,000 tokens, a maximum output of 128,000 tokens and a knowledge cutoff of April 30, 2026. It accepts text and image input and returns text. The reasoning.effort parameter accepts low, medium (the default), high, xhigh and max; none and minimal are not supported. OpenAI directs developers to the Responses API for tool calling, while Chat Completions works without tool calling. Prompts above 272,000 input tokens are charged at twice the input and cache rates and 1.5 times the output rate for the full request. Batch and Flex are 50 percent cheaper than standard, and Fast mode costs twice standard. US and EU data residency are supported, but Fast mode is unavailable with EU residency. Fine-tuning is not supported. OpenAI also plans a GPT-6.1 Sol Ultrafast option in Codex within days, promising up to 8x faster token generation than standard speed.

Among competing first-party list prices verified on September 30, 2026, Claude Sonnet 5.5 matches Sol's $2 input and $10 output rates, but Sol's cached input costs half as much. OpenAI's benchmarks compare Sol against Claude Opus 5.5 rather than Sonnet 5.5. Gemini 3.1 Pro Preview matches Sol on input price, charges $12 for output and remains in preview. List prices are not direct cost comparisons: Anthropic notes that its newer tokenizer produces roughly 30 percent more tokens for the same text.

Editor's Summary

OpenAI released GPT-6.1 Sol, a closed-weight hosted model priced at $2 per million input tokens, $10 per million output tokens and $0.10 per million cached input tokens, or roughly one-fifth of GPT-6 Astra's standard rates. OpenAI's vendor-reported benchmarks place it near Astra on agentic coding and computer use while exceeding Claude Opus 5.5 on several professional and automation evaluations at lower cost per task. Astra remains the recommended option for the hardest scientific work, and list prices are not directly comparable across vendors because of differences in tokenization.