Model comparison
DeepSeek-V4-Pro vs GPT-6 Sol: API costs for agent work
DeepSeek-V4-Pro costs $0.95 for a cached coding session against $2.10 on GPT-6 Sol. How the gap forms, and what DeepSeek's Codex support does and doesn't mean.
· Prices as of September 28, 2026
DeepSeek-V4-Pro
DeepSeek · Released August 13, 2026
DeepSeek's agent-focused large model, released in general availability in August 2026 with support for OpenAI's Responses API and Codex.
DeepSeek-V4-Pro facts and comparisonsGPT-6 Sol
OpenAI · Released September 22, 2026
The mid-priced GPT-6 model, which OpenAI pitches for complex coding and agent workflows and which the Codex docs recommend for complex coding.
GPT-6 Sol facts and comparisons
The short answer
DeepSeek-V4-Pro costs $0.95 for the example agentic coding session against $2.10 on GPT-6 Sol, a 2.2x gap that comes mostly from cache pricing: DeepSeek lists no write fee and charges $0.044 per million for a hit, against $0.20 on Sol. GPT-6 Sol is the model the Codex docs recommend for complex coding, while DeepSeek says it adapted V4-Pro for Codex through native Responses API support. Pick GPT-6 Sol to stay on OpenAI's own stack, and DeepSeek-V4-Pro for lower API costs or open weights you can host.
Choose DeepSeek-V4-Pro if
- You want the lower rate on every line: $1.32 input and $3.96 output per million at DeepSeek's peak, against $2 and $10 on Sol.
- Your harness already speaks OpenAI's Responses API, which DeepSeek says V4-Pro supports natively.
- You want MIT-licensed open weights that you can run on your own hardware.
- Your agent writes long outputs, up to 384K tokens per response against Sol's 128K.
Choose GPT-6 Sol if
- Codex is where you work, and you want the model its docs recommend for complex coding.
- You are still on GPT-5.6 Sol, GPT-5.6 Terra, or GPT-5.4, all of which Codex suggests replacing with GPT-6 Sol.
- GitHub Copilot is part of your workflow, and it offers GPT-6 Sol but not DeepSeek-V4-Pro.
- You want OpenAI's cache controls: up to four explicit breakpoints and a prefix that stays reusable for at least 30 minutes after its last use.
Side by side
Specs and prices
| Fact | DeepSeek-V4-Pro | GPT-6 Sol |
|---|---|---|
| Maker | DeepSeek | OpenAI |
| API model id | deepseek-v4-pro | gpt-6-sol |
| Released | August 13, 2026 | September 22, 2026 |
| Status | Current | Current |
| Context window | 1M tokens | 1.05M tokens |
| Max output | 384K tokens | 128K tokens |
| Open weights | Yes | No |
| Input, per 1M tokens | $1.32 | $2 |
| Cache hit, per 1M | $0.044 | $0.20 |
| Cache write, per 1M | $1.32 (same as input) | $2.50 |
| Output, per 1M tokens | $3.96 | $10 |
| Runs in | OpenCode and OpenRouter | Codex, OpenCode, OpenRouter, and GitHub Copilot |
Standard API rates in US dollars, as published by each maker on September 28, 2026. Batch and priority tiers, taxes, and subscription plans are not included. DeepSeek-V4-Pro: Prices are DeepSeek's peak rates. Off-peak hours cost 50% less: peak is 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, excluding Chinese public holidays. GPT-6 Sol: Requests over 272K input tokens cost 2x for input and cache and 1.5x for output, for the whole request.
Cost
What the same work costs
The same token counts, priced at each model’s published rates. Cached tokens are billed at each maker’s cache prices, so the session shows what caching is worth on each model.
Real sessions differ: the two models count the same code as different numbers of tokens, and reasoning settings change how much each one writes. Your own history is the real test.
| Workload | DeepSeek-V4-Pro | GPT-6 Sol |
|---|---|---|
| Agentic coding session, 100K input, 400K written to cache, 2M read from cache, 50K output | $0.95 | $2.10 |
| Large one-off review, 150K input with no cache hits, 10K output | $0.24 | $0.40 |
| Output-heavy generation, 30K input, 80K output | $0.36 | $0.86 |
| A month of sessions, 110 sessions: 5 a day, 22 working days | $104.06 | $231.00 |
| Where the session’s cost goes | ||
| Cache writes | $0.53 | $1.00 |
| Cache reads | $0.09 | $0.40 |
| Uncached input | $0.13 | $0.20 |
| Output | $0.20 | $0.50 |
- caching saves on the session with DeepSeek-V4-Pro (73%)
- $2.55
- caching saves on the session with GPT-6 Sol (62%)
- $3.40
DeepSeek-V4-Pro and Codex: what DeepSeek says
Codex is OpenAI's own coding agent, and its docs recommend GPT-6 Sol for complex coding. OpenAI describes Sol as "Built to power complex coding and agentic workflows," and sets its reasoning effort to medium by default, both in the API and inside Codex.
DeepSeek is aiming at the same workflow from outside. It says the general availability release of DeepSeek-V4-Pro in August 2026 added native support for OpenAI's Responses API, adapted for Codex with one-click setup. That is DeepSeek's description of its own integration. The Codex docs' recommendation for complex coding is GPT-6 Sol.
Cost tracking follows the provider, not the tool. EveryToken prices GPT-6 Sol at OpenAI's rates from Codex or OpenCode history. It prices DeepSeek-V4-Pro when requests go through OpenRouter, from OpenRouter's catalog, and does not price it when you call DeepSeek's own API, even from inside Codex.
Why GPT-6 Sol costs 2.2x as much per session
On input the two are close: $1.32 per million on DeepSeek-V4-Pro at peak against $2 on GPT-6 Sol, 1.5x. The large one-off review, which is almost all input, shows 1.7x, $0.24 against $0.40. Output is further apart, $3.96 against $10, which makes the output-heavy generation 2.4x, $0.36 against $0.86.
Caching widens the gap on the session. A V4-Pro cache hit costs $0.044 per million, 3.3% of its input rate, while Sol follows OpenAI's 10% rule at $0.20. OpenAI charges 1.25x input to write Sol's cache, $2.50 per million, and DeepSeek lists no write fee, so the session's 400K written tokens cost $0.53 on V4-Pro against $1.00 on Sol.
Add it up and the session comes to $0.95 on DeepSeek-V4-Pro and $2.10 on GPT-6 Sol. Over 110 sessions a month that is $104.06 against $231.00, a $126.94 difference, and caching trims 73% off V4-Pro's uncached cost against 62% off Sol's.
Peak hours, Sol's 272K line, and other limits
DeepSeek's figures here are peak rates, charged from 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, excluding Chinese public holidays. Every other hour of the week costs 50% less on V4-Pro. Sol has a size rule instead: past 272K input tokens, OpenAI bills the entire request at 2x for input and cache and 1.5x for output.
Sol takes up to 922K input tokens of its 1.05M context window and writes up to 128K. DeepSeek-V4-Pro takes 1M and writes up to 384K. Each request in the example session stays below 200K, so Sol's 272K rule plays no part in these tables.
Two more notes from DeepSeek. The current snapshot is DeepSeek-V4-Pro-0813, and service continues with billing unchanged until a V4.1 Pro model arrives. DeepSeek also says its newer DeepSeek-V4.1-Flash is ahead of V4-Pro on performance, cost, speed, and total runtime in tests by several parties.
Both models are on OpenRouter. There, a request for DeepSeek's open-weight model goes to one of several providers, each with its own price that can differ from DeepSeek's; the tables use each maker's own API price.
Prompt caching
How each maker bills cached tokens
DeepSeek
DeepSeek's disk cache is on by default for every account, with no code changes. DeepSeek lists no fee for writing the cache.
A cache hit costs $0.006 per million tokens on DeepSeek-V4.1-Flash and $0.044 on DeepSeek-V4-Pro at peak rates, and off-peak hours cost 50% less.
Source: DeepSeek API: Context caching
OpenAI
Prompt caching is on by default. From GPT-5.6 on you can also mark up to four explicit cache breakpoints, while GPT-5.5 and earlier cache automatically only.
From GPT-5.6 on, a cache write costs 1.25x the uncached input price and a cache hit costs 0.1x. GPT-5.5 and earlier add no charge for writing the cache: written tokens are billed as ordinary input, and a hit costs 0.1x on the models compared here.
On GPT-5.6 and later, a cached prefix stays reusable for at least 30 minutes after its last use, and caching starts at 1,024 input tokens.
Source: OpenAI: Prompt caching
Your own numbers
See what DeepSeek-V4-Pro and GPT-6 Sol really cost you.
everyaitoken reads your OpenRouter, Codex, and OpenCode history on your Mac and prices every request at API rates, with what caching saved or cost. $9 once.
FAQ
Questions
Can I use DeepSeek-V4-Pro in Codex?
DeepSeek says V4-Pro supports OpenAI's Responses API natively and is adapted for Codex with one-click setup. Codex itself is OpenAI's agent, and its docs recommend GPT-6 Sol for complex coding.
How much less does DeepSeek-V4-Pro cost than GPT-6 Sol?
The example cached session costs $0.95 against $2.10, a 2.2x gap, at DeepSeek's peak rates. Uncached work is closer: a large one-off review costs $0.24 against $0.40.
What reasoning effort does each model use?
OpenAI starts Sol at medium, whether you call it through the API or run it in Codex. DeepSeek offers V4-Pro at low, high, or max. Raising effort tends to raise output, which is the pricier side of both rate cards.
Does GPT-6 Sol charge more for long prompts?
Yes. Once a Sol request passes 272K input tokens, the whole request is billed at 2x for input and cache and 1.5x for output. Sol accepts up to 922K input tokens in total.
Sources
- DeepSeek API: Models and pricing
- DeepSeek API: Change log
- DeepSeek: V4 Pro release
- OpenRouter: DeepSeek-V4-Pro-0813
- OpenCode docs: Zen
- OpenAI: API pricing
- OpenAI docs: GPT-6 Sol
- OpenAI: Introducing GPT-6 Sol and Luna
- OpenAI: API changelog
- Codex docs: Models
- OpenRouter: GPT-6 Sol
- OpenCode docs: Zen
- GitHub Docs: Supported AI models in Copilot
- DeepSeek API: Context caching
- OpenAI: Prompt caching