BALANCED GPT-6 TIERGPT-6 FAMILY1.05M CONTEXTCACHE READS AT 5%TEXT + VISION
GPT-6.1 Sol API Pricing
GPT-6.1 Sol is OpenAI's balanced GPT-6 model, released September 29, 2026 at $2.00/M input, $0.10/M cached input and $10.00/M output. That is the GPT-6 Sol card from a week earlier with one change: the cache read falls from $0.20 to $0.10, 5% of input. OpenAI now lists 6.1 Sol in 6 Sol's place between Astra and Luna, and 6 Sol stays on the price list. Prompts over 272K input tokens bill on the long-context card at $4.00/M input and $15.00/M output.
Input - per 1M tokens
$2.00/M
Short context up to 272K input standard
Output - per 1M tokens
$10.00/M
Long context output is $15/M standard
Cached input - 95% off
$0.10/M
Prompt cache read price -95%
Effective - agentic blend
$1.29/M
92/8 split - 82% cache
§ 01 / TERMINAL
GPT-6.1 Sol API cost calculator.
Live calculator pre-loaded with GPT-6.1 Sol's standard short-context rates. Reasoning effort runs from low to max - none and minimal are not supported on this model - and reasoning tokens bill as output, so put them in the reasoning field. Cross 272K input tokens in one prompt and the long-context card applies instead.
$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
GPT-6.1 Sol cost on real workloads.
CODING AGENT
Repo-wide architecture migration
$0.184/task
RESEARCH
Multi-step diligence brief
$0.072/brief
LONG DOC ANALYSIS
Reading 150-page contracts
$0.208/doc
ASSISTANT
Executive workflow turn
$0.00606/turn
Exact · o200k_base Auto-counts as you type
Counted with the o200k_base BPE encoding — the same tokenizer the API uses — entirely in your browser. The encoder loads on your first keystroke (~1 MB, one time); your text never leaves this page.
Characters 593
Words 98
Tokens (exact) 147 tokens
Cost as input · uncached $0.00029 USD
Cost as output · uncached $0.00147 USD
Cost as cached input $0.00001 USD
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| GPT-6.1 Sol Current | $2.00 cache $0.10 | $10.00 | $1.29 agentic 92/8 | 1.05M | Balanced coding and professional work |
| GPT-6 Astra | $10.00 cache $1.00 | $50.00 | $6.82 +5.3× | 1.05M | OpenAI's top-of-stack reasoning tier |
| GPT-6 Luna | $0.10 cache $0.01 | $0.50 | $0.0682 −19× | 1.05M | High-volume routing tier |
| GPT-6 Sol | $2.00 cache $0.20 | $10.00 | $1.36 +6% | 1.05M | The previous Sol, with the none effort |
| GPT-5.6 Terra | $2.00 cache $0.20 | $12.00 | $1.52 +18% | 1.05M | Balanced GPT-5.6 agents |
| Claude Sonnet 5.5 | $2.00 cache $0.10 | $10.00 | $1.29 same | 1M | Anthropic's model at the same list price |
| Grok 4.7 | $2.00 cache $0.50 | $6.00 | $1.19 −8% | 500K | Cheaper frontier coding model |
| Gemini 3.8 Flash | $0.75 cache $0.075 | $3.75 | $0.481 −2.7× · promo | 1M | Multimodal high-volume alternative |
Compare siblings
GPT-6.1 Sol has published scores on 9 benchmarks we track, and shares first place on ARC-AGI-1. 4 of these scores come from an independent leaderboard; the rest are the lab's own reports.
| Benchmark | Score | Rank | Leader | Source |
|---|---|---|---|---|
| ARC-AGI-1 Reasoning | 98.5% ARC Prize verified, best result | #1 of 99 | Tied for first with 5 more | Official leaderboard ARC Prize Foundation |
| ARC-AGI-2 Reasoning | 94.2% ARC Prize verified, best result | #2 of 105 | GPT-6 Astra · 95% | Official leaderboard ARC Prize Foundation |
| ARC-AGI-3 Reasoning · Standard harness | 52.7% ARC Prize verified, best result | #2 of 24 | GPT-6 Astra · 62.7% | Official leaderboard ARC Prize Foundation |
| Terminal-Bench 4.0 Agents and computer use | 58.2% max effort, Codex harness, 330 trials, ±3.1 pts (95% CI) | #5 of 34 | Kimi K3.1 Pro · 65.7% | Official leaderboard Terminal-Bench (Stanford, Laude Institute) |
| DeepSWE Coding | 75.2% high effort | #4 of 27 | Gemini 4 Argon · 77.9% | Vendor-reported OpenAI |
| OSWorld 2 Agents and computer use · 2.0, offline subset, partial score | 71.4% max effort | #2 of 7 | GPT-6 Astra · 73.5% | Vendor-reported OpenAI |
| Terminal-Bench-Science AI research | 57% max effort | #7 of 12 | Claude Mythos 5.2 · 75.3% | Vendor-reported OpenAI |
| AutomationBench Agents and computer use · Main set (as reported) | 36.1% max effort, v1.0.6 | #13 of 22 | MiniMax M3.1 Flash · 62.8% | Vendor-reported OpenAI |
| GDP.pdf Vision and documents · Default setting | 32% high effort | #5 of 7 | Gemini 3.8 Flash · 35% | Vendor-reported OpenAI |
GPT-6.1 Sol API pricing FAQ.
GPT-6.1 Sol pricing questions - the standard card, what changed from GPT-6 Sol, and where it sits against Astra and Claude Sonnet 5.5.
Q · 01 How much does GPT-6.1 Sol cost? +
OpenAI lists
gpt-6.1-sol at $2.00/M input, $0.10/M cached input, $2.50/M cache writes and $10.00/M output for prompts up to 272K input tokens. Under our standard blend (92% input, 82% cache hits) that works out to $1.29/M.Q · 02 What changed from GPT-6 Sol? +
On price, only the cache read:
$0.10/M against $0.20/M, so the same workload blends to $1.29/M instead of $1.36/M. Input, cache writes and output are identical. Two behaviours differ: 6.1 Sol does not accept the none or minimal reasoning effort, and it needs the Responses API for tool calling, where 6 Sol can call functions in Chat Completions with reasoning off. OpenAI now names 6.1 Sol in its model list and points gpt-6-sol users to migration guidance; GPT-6 Sol remains on the price list.Q · 03 When does long-context pricing apply? +
When a prompt exceeds
272K input tokens. The whole request then bills at $4.00/M input, $0.20/M cached input and $15.00/M output - input and cache rates double, output rises by half.Q · 04 How much cheaper are Batch and Flex? +
Both are half the standard card:
$1.00/M input and $5.00/M output, with cached input at $0.05/M. Fast mode, the tier OpenAI renamed from Priority on July 30, 2026, runs at double standard, $4/M and $20/M.Q · 05 How does it compare with GPT-6 Astra? +
GPT-6 Astra lists at
$10.00 / $50.00, five times 6.1 Sol on both sides, and blends to $6.82/M. OpenAI's pitch for 6.1 Sol is "near-Astra performance at a lower cost" and its advice is to run both on your own tasks before choosing.Q · 06 How does it compare with Claude Sonnet 5.5? +
Claude Sonnet 5.5 has the same input and output sticker,
$2.00 / $10.00, and the same cache-write rate. Since October 7, 2026 the cache read matches too: Anthropic halved Sonnet 5.5's to $0.10/M, the same as $0.10/M here, so the cache-heavy agent blends come to $1.29/M and $1.29/M. What still separates them is the long-context card above 272K input tokens, which Sonnet 5.5 does not have.Q · 07 What is GPT-6.1 Sol Pro? +
Not a separate model. Some resellers list a "GPT-6.1 Sol Pro", which is
gpt-6.1-sol called with reasoning.mode set to pro. OpenAI bills pro mode at the model's standard token rates; the model simply does more work, so the same request uses more tokens and costs more.Q · 08 What is the context window? +
1.05M tokens, of which up to 922K can be input, with 128K max output and a knowledge cutoff of April 30, 2026. It takes text and image input and returns text, on the Responses and Chat Completions APIs.Q · 09 Does regional processing cost more? +
Yes, 10% on regional data-residency endpoints. GPT-6.1 Sol supports US and EU data residency, but Fast mode is not available with EU residency.