OpenAI

gpt-5.6-sol

Made by OpenAI. Released 2026-07-09. Accessibility not recorded. Cheapest standard rate we hold: $20 per million output tokens, at short context.

What it costs

TierKindInOutCached inFrom
batchtext · long$4$15$0.40seller
batchtext · short$2$10$0.20seller
fasttext · long$16$60$1.60seller
fasttext · short$8$40$0.80seller
flextext · long$4$15$0.40seller
flextext · short$2$10$0.20seller
standardtext · long$8$30$0.80seller
standardtext · short$4$20$0.40seller

per million tokens, US dollars. * marks a rate from a single registry or two that disagree — the best available number rather than a confirmed one. Every other row is confirmed against a seller or a licensed index.

What it scores, by house

BenchmarkHouseScoreEvaluatedScaffold
chess puzzlesEpoch AI0.3±0.042026-08-07low
chess puzzlesEpoch AI0.1±0.032026-08-07none
chess puzzlesEpoch AI0.6±0.052026-07-10promax
chess puzzlesEpoch AI0.6±0.052026-07-09max
ebr benchEpoch AI0.4±0.042026-09-04max
frontiermath erdosEpoch AI0.0±0.002026-08-28max
frontiermath tier 4 v2Epoch AI0.8±0.062026-07-09max
frontiermath tier 4 v2Epoch AI0.8±0.062026-07-09promax
frontiermath tiers 1 3 v2Epoch AI0.9±0.022026-07-09max
gpqa diamondEpoch AI0.8±0.032026-08-07none
gpqa diamondEpoch AI0.9±0.022026-08-07low
gpqa diamondEpoch AI0.9±0.022026-07-09max
mirrorcodeEpoch AI0.2±0.092026-08-10high
mystery game puzzlesEpoch AI0.3±0.052026-08-27none
mystery game puzzlesEpoch AI0.3±0.042026-08-27low
mystery game puzzlesEpoch AI0.6±0.052026-07-28max
otis mock aime 2024 2025near its ceilingEpoch AI0.7±0.072026-08-07none
otis mock aime 2024 2025near its ceilingEpoch AI1.0±0.032026-08-07low
otis mock aime 2024 2025near its ceilingEpoch AI1.0±0.002026-07-09max
simpleqa verifiedEpoch AI0.7±0.012026-08-10max
ale benchcollected by Epoch AI2177max
apex agentscollected by Epoch AI0.6max
apex agentscollected by Epoch AI0.6promax
apex agentscollected by Epoch AI0.5unknown
apex agentscollected by Epoch AI0.5xhigh
arc aginear its ceilingcollected by Epoch AI0.9medium
arc aginear its ceilingcollected by Epoch AI1.0xhigh
arc aginear its ceilingcollected by Epoch AI1.0high
arc aginear its ceilingcollected by Epoch AI1.0max
arc aginear its ceilingcollected by Epoch AI0.7low
arc agi 2collected by Epoch AI0.9xhigh
arc agi 2collected by Epoch AI0.9max
arc agi 2collected by Epoch AI0.7medium
arc agi 2collected by Epoch AI0.4low
arc agi 2collected by Epoch AI0.9high
blueprint bench 2collected by Epoch AI0.3unknown
btf3collected by Epoch AI0.1FutureSearch agent
critptcollected by Epoch AI0.1low
critptcollected by Epoch AI0.1none
critptcollected by Epoch AI0.2medium
critptcollected by Epoch AI0.3max
critptcollected by Epoch AI0.3high
critptcollected by Epoch AI0.3xhigh
cursorbenchcollected by Epoch AI0.6medium
cursorbenchcollected by Epoch AI0.7max
cursorbenchcollected by Epoch AI0.6high
cursorbenchcollected by Epoch AI0.5low
cursorbenchcollected by Epoch AI0.6xhigh
deepswecollected by Epoch AI0.7mini-swe-agent
dtbenchnear its ceilingcollected by Epoch AI1.0max
dtbenchnear its ceilingcollected by Epoch AI1.0promax
enigma evalcollected by Epoch AI0.4high
frontiercodecollected by Epoch AI0.5codex
frontierswecollected by Epoch AI0.3proximus
gbaevalcollected by Epoch AI0.5unknown
gdp pdfcollected by Epoch AI0.3max
gdp pdfcollected by Epoch AI0.3unknown
lmcacollected by Epoch AI59.2promax
lmcacollected by Epoch AI58.4max
osworld 2collected by Epoch AI0.3max
posttrainbenchcollected by Epoch AI0.4Codex CLI
proofbenchcollected by Epoch AI0.8max
scicodecollected by Epoch AI0.6medium
scicodecollected by Epoch AI0.6high
scicodecollected by Epoch AI0.6low
scicodecollected by Epoch AI0.5none
scicodecollected by Epoch AI0.6max
scicodecollected by Epoch AI0.6xhigh
simplebenchcollected by Epoch AI0.6unknown
simplebenchcollected by Epoch AI0.7prounknown
simplebenchcollected by Epoch AI0.6xhigh
surface evolver benchcollected by Epoch AI0.9xhigh
vending bench 2collected by Epoch AI9619unknown
webdev arenacollected by Epoch AI1619xhigh
weirdmlcollected by Epoch AI0.9max
weirdmlcollected by Epoch AI0.9promax
weirdmlcollected by Epoch AI0.9high

These scores are not comparable between houses and are not added up. A score is dated by when it was evaluated, not when the model was released, and a score without its scaffold is not a measurement — which is why both are printed.

Every rate we hold · Why there is no single best model · Everything from OpenAI