GPT-6.1 Sol Launches with Astra-level Intelligence at One-fifth the Price

OpenAI launched GPT-6.1 Sol on 09-30, pitching it as "Astra-level intelligence at roughly 5x cheaper." Per third-party evaluator Artificial Analysis, the model replaced GPT-6 Sol within just 7 days of release, scoring only 1 point below GPT-6 Astra on the Intelligence Index v4.3.2. The current takeaway: it leads clearly on price-performance and efficiency, but absolute intelligence still trails Anthropic's Opus 5.5. It matters because it could reshape developers' model selection and cost structures.

Confirmed

• Artificial Analysis published a GPT-6.1 Sol evaluation page: the five variants (max/xhigh/high/medium/low) score 42–52, up across all reasoning tiers versus the previous generation (low 42 vs 34 before, medium 48 vs 40, etc.), with per-question cost differing up to 5.5x between tiers.
• Pricing is $2/$10 per million tokens, with the cached-read discount raised from 90% to even higher; per comparison data cited by Reddit user that90sguy, GPT-6 Sol costs about half of Sonnet 5.5, a third of 5.6 Sol, and a fifth of Opus 5.5.
• Artificial Analysis also released Intelligence Index v4.3.2 details, covering ten benchmarks including AA-Briefcase, GDPval-AA, AutomationBench, and Terminal-B; in the same update, Claude Sonnet 5.5 surged to second place on the index.

Unconfirmed

• HLE-Diamond scores shared by haider1 (53.2% overall, 67.4% on the pure-reasoning subset, using only 4k tokens) are reportedly from GPT-6.1 sol but unverified officially.
• charliermarsh reshared blader's internal cfo.ai evals: previously Opus 5.5 was roughly 2x as efficient as GPT-6-sol, and after the 6.1-sol update it pulled ahead to about 3x; these are personal internal data.
• In an npew post, Gabriel, an engineering lead at an Energy company, claimed that in internal enterprise evals GPT 6.1 far exceeded Opus 5.5 in completion rate at 40% lower cost and 2x the speed—a single-source claim.

Why it matters

• Multiple evaluators (rhiever, randalolson, danielmac8) drawing on Artificial Analysis data converge on the same conclusion: Opus 5.5 still leads on the intelligence index, but GPT-6.1 Sol pushes the intelligence-per-dollar Pareto frontier meaningfully outward, making it the pick when optimizing intelligence per dollar.
• dkundel reshared Theo's Terminal Bench 4 self-test: GPT-6.1 Sol outperforms Opus 5.5 at roughly 1/30 the price.
• Community sentiment isn't unanimous: kimmonismus finds it slightly slower but highly efficient, competitive with Sonnet 5.5, while worrying about subscription-plan changes, still preferring Opus 5.5, and predicting the next few major releases may come from China.
• BorisMPower noted the Sol version shipped on schedule, in contrast to earlier reports that GPT-6.1 Astra was pulled over safety concerns.

2026-09-29 ~ 2026-10-01 · 32 related posts

Full story(15 episodes)→

• Episode 1: OpenAI Launches GPT-6 Sol and Luna at Half Price; Big Hallucination Gains but Not a Clean Sweep(2026-09-23, 92 posts)
• Episode 2: GPT-6 Sol/Luna launch splits opinion: half-price but dumber?(2026-09-23, 9 posts)
• Episode 3: LMArena Opens GPT-6 Sol and Luna for Public Testing(2026-09-23, 2 posts)
• Episode 4: Four New Models Reshape the Cost-Performance Frontier, Opus 5.5 Tops Intelligence Index(2026-09-23, 2 posts)
• Episode 5: OpenAI and Anthropic Slash Prices Within 90 Minutes, Kicking Off a New Model Price War(2026-09-23, 7 posts)
• Episode 6: Tests Show GPT-6 Sol Scales Nearly Linearly With Effort(2026-09-23, 3 posts)
• Episode 7: GPT-6 Astra Tops Bug Hunt Bench While Sol Version Regresses(2026-09-24, 2 posts)
• Episode 8: Leaks Map GPT-6 Codenames to GPT-5.6 Versions(2026-09-24, 3 posts)
• Episode 9: Bug Hunt Bench Ranks Frontier Coding Models on 105 Real Bugs with 200x Cost Gap(2026-09-24, 3 posts)
• Episode 10: GPT-6 Sol and Luna Launch With Aggressive API Pricing(2026-09-25, 2 posts)
• Episode 11: OpenAI's efficient GPT-6 models seen as freeing compute for bigger frontier models(2026-09-25, 2 posts)
• Episode 12: GPT-6 Sol Underestimated: 1.6x Robot Task Gains at 47% Lower Cost(2026-09-25, 2 posts)
• Episode 13: GPT-6.1 Sol Launches with Astra-level Intelligence at One-fifth the Price(2026-09-29, 32 posts)
• Episode 14: GPT-6 Astra Ultrafast Reportedly 8x Faster in Codex(2026-09-30, 2 posts)
• Episode 15: GPT-6.1 Sol Delivers Near-Astra Quality at Fraction of Cost, Users Report(2026-09-30, 2 posts)

• Databricks evals: Opus 5.5 cuts task costs 20%, GPT-6 Luna is 20x cheaper — pwendell · 2026-09-29
• [source] GPT-6.1 Sol Launches, Promising Astra-Level Intelligence at 5x Cheaper — BorisMPower · 2026-09-30
• GPT-6.1 Sol matches GPT-6 Astra at ~1/5 cost, ultrafast version targets 300 tok/s — haider1 · 2026-09-30
• Early-access tester claims GPT-6.1 Sol matches GPT-6 Astra, faster and much cheaper — DeryaTR_ · 2026-09-30
• GPT-6.1 Sol computer use nearly matches GPT-6 Astra at 1/7 the cost on OSWorld 2.0 — haider1 · 2026-09-30
• GPT 6.1 hits Artificial Analysis Intelligence Index gains across all effort tiers — Background-Web-6312 · 2026-09-30
• Theo's own Terminal Bench 4 runs: GPT-6.1 Sol beats Opus 5.5 at ~1/30th the price — dkundel · 2026-09-30
• Sol 6.1 matches Astra on benchmarks at one-fifth the price, says founder — bindureddy · 2026-09-30
• GPT-6.1 Sol shows solid index gains, now neck-and-neck with GPT-6 Astra and Fable 5.1 — haider1 · 2026-09-30
• GPT-6.1-Sol shifts the intelligence-per-dollar frontier, but Opus 5.5 still tops raw scores — randal_olson · 2026-09-30
• [source] GPT-6.1 Sol replaces GPT-6 Sol in 7 days, scoring 1 point below Astra at quarter the cost — ArtificialAnlys · 2026-09-30
• Opus 5.5 still tops AA intelligence, but GPT-6.1 Sol shifts the cost frontier — rhiever · 2026-09-30
• Reddit user defends GPT-6.1 Sol: a fifth the price of Opus 5.5 with better intelligence per dollar — that_90s_guy · 2026-09-30
• Internal evals put GPT 6.1 Sol on top: beats Claude Opus 5.5 at 40% cost, 2x speed — emilahlback · 2026-09-30
• Internal evals: GPT 6.1 Sol beats Claude Opus 5.5 on knowledge work at 40% of the cost — emilahlback · 2026-09-30
• Reddit user argues GPT-6/6.1 Sol's price — 1/2 to 1/5 of rivals — outweighs bench complaints — that_90s_guy · 2026-09-30
• GPT-6.1 Sol Trails Opus and Sonnet 5.5 on Intelligence but Is Pareto-Optimal on Efficiency — daniel_mac8 · 2026-09-30
• Writing benchmark: GPT-6.1 Sol regresses 153 Elo below GPT-6 Sol while costing more — OnlyProggingForFun · 2026-09-30
• Artificial Analysis breaks down individual evals in Intelligence Index v4.3.2 — ArtificialAnlys · 2026-09-30
• [source] GPT-6.1 Sol benchmarks: 5 tiers score 42-52, cost-per-task varies 5.5x — ArtificialAnlys · 2026-09-30
• GPT-6.1 Sol called a huge, affordable leap that rivals Sonnet 5.5 — kimmonismus · 2026-09-30
• GPT-6.1 sol reportedly tops pure reasoning on HLE-Diamond at 67.4% using just 4k tokens — haider1 · 2026-09-30
• Enterprise evals: GPT-6.1 beats Opus 5.5 with 2x speed and 40% lower price — npew · 2026-09-30
• User claims GPT-6.1 Sol nearly matches Opus 5.5 at a fraction of the price — shaunralston · 2026-09-30
• Internal evals: GPT-6.1-sol hits 3x the efficiency of Opus 5.5 in surprise point release — charliermarsh · 2026-09-30
• Artificial Analysis updates: GPT-6.1 Sol tops GPT-6 Sol in 7 days, Claude Sonnet 5.5 hits #2 — ArtificialAnlys · 2026-09-30
• GPT-6.1 Sol takes #1 on MathArena, cheaper and better than Astra — airesearch12 · 2026-09-30
• GPT-6.1 Sol hands-on: efficient but slow, and subscription changes worry users — kimmonismus · 2026-09-30
• gpt-6.1 sol max claims: cheaper than astra low, nearly as smart as astra max — pvncher · 2026-10-01

3 near-duplicate retellings: pwendell · rhiever · emilahlback