OpenAI Releases GPT-6.1 Sol: Near-Astra Coding and Computer Use at One-Fifth of Astra's Token Price

This week, OpenAI released GPT-6.1 Sol. It upgrades GPT-6 Sol, the mid-tier model in the GPT-6 family. OpenAI’s claim is specific: near-Astra results on agentic coding, computer use, and professional work. The price is one-fifth of GPT-6 Astra’s standard input and output rates. Cached input drops to $0.10 per million tokens, 50% below GPT-6 Sol.

Is it deployable? Yes, as a hosted model. It is live today in the OpenAI API as gpt-6.1-sol, and in ChatGPT Work and Codex.

Elon Musk And Palmer Luckey Will Advise The Government On The Future Of Warfare

Google DeepMind Unveils Gemini 4 Argon with 1M Output Tokens for Coding, Knowledge Work and Cyber Defense

Where Sol Sits in the GPT-6 Lineup

OpenAI now ships 3 GPT-6 tiers. GPT-6 Astra costs $10 input, $50 output, and $1 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.

The cached price is quite important most for agents. Agents resend the same system prompt, tool schemas, and history on every step. Cached reads now cost 5% of the uncached input rate, down from 10% on GPT-6 Sol.

Benchmark Results

All figures below are vendor-reported in OpenAI’s launch post. OpenAI states that competitor numbers came from public reports.

Coding: On DeepSWE v1.1, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth of the cost. It beats GPT-6 Sol’s best score by 6.4 percentage points, at a lower reasoning effort.
Professional work: On GDP.pdf, which tests answers over complex professional PDFs, Sol beats Claude Opus 5.5 with fallbacks. It does so 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. That result comes at roughly one-third the cost.
Computer use: On the OSWorld 2.0 offline set, Sol gains 7 points over GPT-6 Sol at maximum effort. It lands within 2.1 points of Astra at roughly one-seventh the cost per task.
Science: On Terminal-Bench Science 0.1, Sol more than doubles GPT-6 Sol’s score at max effort. Average cost per task is $5.47, versus $23.21 for Opus 5.5 and $23.80 for Astra. Astra still leads at 68.1%, and OpenAI recommends it for the hardest research.
Factuality: At low effort, the share of responses with a factual error falls from 11.4% to 7.7%. That is a reduction of about 32%. The eval uses deliberately difficult conversations where users had flagged earlier model errors.

API Details Developers Need

From the GPT-6.1 Sol model page:

Context window of 1,050,000 tokens, 128,000 max output tokens, and an April 30, 2026 knowledge cutoff.
Text and image input; text output.
reasoning.effort accepts low, medium (default), high, xhigh, and max. The none and minimal settings are not supported.
Use the Responses API for tool calling. Chat Completions works without tool calling.
Prompts above 272K input tokens cost 2x input and cache rates, and 1.5x output, for the full request.
Batch and Flex are 50% cheaper. Fast mode costs 2x standard.
US and EU data residency are supported. 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. It promises up to 8x faster token generation than standard speed.

Interactive Explainer

=1?(n%1===0?n:n.toFixed(1))+’K’:(n*1000)+”} function money(x){return x‘; d.dataset.c=cached;d.dataset.f=fresh;loop.appendChild(d)} } var playing=false; function play(){ if(playing)return;playing=true; var cells=$(‘loop’).children,k=0; Array.prototype.forEach.call(cells,function(c){c.children[0].style.height=”0″;c.children[1].style.height=”0″;c.classList.remove(‘active’)}); (function step(){ if(k>0)cells[k-1].classList.remove(‘active’); if(k>=cells.length){playing=false;return} var c=cells[k];c.classList.add(‘active’); var cached=+c.dataset.c,fresh=+c.dataset.f; c.children[1].style.height=cached+’%’; c.children[0].style.bottom=cached+’%’;c.children[0].style.height=Math.max(0,Math.min(fresh,100-cached))+’%’; k++;setTimeout(step,260); })(); } [‘S’,’C’,’H’,’O’].forEach(function(id){$(id).addEventListener(‘input’,calc)}); $(‘play’).addEventListener(‘click’,play); calc();setTimeout(play,500); // —- Pane 2: tiers —- var T={astra:{i:10,o:50,c:1},sol:{i:2,o:10,c:0.10},luna:{i:0.10,o:0.50,c:0.01}}; var key=’i’,sel=”sol”; var names={i:’input’,o:’output’,c:’cached input’}; function pf(x){return x>=1?’$’+x:’$’+x.toFixed(2)} function drawTiers(){ var mx=T.astra[key]; $(‘pa’).textContent=pf(T.astra[key]);$(‘ps’).textContent=pf(T.sol[key]);$(‘pl’).textContent=pf(T.luna[key]); [‘a’,’s’,’l’].forEach(function(x){$(‘t’+x).style.width=”0″}); setTimeout(function(){ $(‘ta’).style.width=”100%”;$(‘ts’).style.width=Math.max(1.5,T.sol[key]/mx*100)+’%’;$(‘tl’).style.width=Math.max(1.5,T.luna[key]/mx*100)+’%’; },40); $(‘tva’).textContent=pf(T.astra[key]);$(‘tvs’).textContent=pf(T.sol[key]);$(‘tvl’).textContent=pf(T.luna[key]); var m=T[sel],label={astra:’GPT-6 Astra’,sol:’GPT-6.1 Sol’,luna:’GPT-6 Luna’}[sel]; var txt=”“+label+’ costs ‘+pf(m.i)+’ input, ‘+pf(m.o)+’ output and ‘+pf(m.c)+’ cached input per 1M tokens.’; if(sel===’sol’)txt+=’ Its ‘+names[key]+’ price is ‘+(T.astra[key]/T.sol[key])+’x lower than Astra.’; if(sel===’astra’)txt+=’ OpenAI still recommends Astra for the hardest scientific research tasks.’; if(sel===’luna’)txt+=’ Luna sits below Sol for high-volume, cost-sensitive work.’; $(‘tdetail’).innerHTML=txt; } root.querySelectorAll(‘#seg button’).forEach(function(b){b.addEventListener(‘click’,function(){ root.querySelectorAll(‘#seg button’).forEach(function(x){x.classList.remove(‘on’)});b.classList.add(‘on’);key=b.dataset.k;drawTiers(); })}); root.querySelectorAll(‘.tier’).forEach(function(t){t.addEventListener(‘click’,function(){ root.querySelectorAll(‘.tier’).forEach(function(x){x.classList.remove(‘sel’)});t.classList.add(‘sel’);sel=t.dataset.m;drawTiers(); })}); // —- Pane 3: results —- function drawCols(){ root.querySelectorAll(‘.vcol’).forEach(function(col){ var mx=+col.dataset.max; col.querySelectorAll(‘.v’).forEach(function(v){v.style.height=”0″}); setTimeout(function(){col.querySelectorAll(‘.v’).forEach(function(v){v.style.height=Math.max(2,(+v.dataset.h/mx)*82)+’%’})},60); }); } var D=[ ‘On DeepSWE v1.1, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth of the cost. It beats GPT-6 Sol\’s best score by 6.4 points at a lower reasoning effort.’, ‘On AutomationBench 1.0.6 (47 tools across business workflows), Sol scores 2.2 points above Claude Opus 5.5 at medium effort, at roughly one-third the cost. It is up 4.8 points over GPT-6 Sol.’, ‘On the OSWorld 2.0 offline set, Sol gains 7 points over GPT-6 Sol at max effort for less than half the cost. It lands within 2.1 points of Astra at roughly one-seventh the cost per task.’, ‘On GDP.pdf (questions over complex professional PDFs), Sol scores higher than Opus 5.5 with fallbacks at less than half the cost per task. It approaches Astra at roughly one-fifth the cost.’, ‘On Terminal-Bench Science 0.1, Sol more than doubles GPT-6 Sol\’s score at max effort. Astra still leads at 68.1% and remains OpenAI\’s pick for the hardest research.’ ]; var chips=root.querySelectorAll(‘.chip’); chips.forEach(function(c){c.addEventListener(‘click’,function(){ chips.forEach(function(x){x.classList.remove(‘on’)});c.classList.add(‘on’); var el=$(‘bdetail’);el.style.opacity=0;setTimeout(function(){el.innerHTML=D[+c.dataset.d];el.style.transition=’opacity .35s’;el.style.opacity=1;post()},150); })}); $(‘bdetail’).innerHTML=D[0]; drawTiers(); setTimeout(post,100); })();

Credit: Source link