SYLIN

GPT Sol

by OpenAI · Model · Language & code models

mixed

Near-top intelligence at a low price, but behind rivals on coding and hard tasks

GPT 6.2 Sol released: less intelligent than Astra but more efficient

AccessAPIPrice$2 in · $10 out per M tokensContext1,050K tokens

Why we say so

Artificial Analysis says 6.2 ties last September's Astra on its Intelligence Index at about 15% of the cost per task, and ARC Prize found it comparable to top models at much lower cost. Simon Willison found its three.js animation weaker than Fable 5.2 and Gemini Carbon, Vals put Astra first and Sol third on Vibe Code Bench, and an HN commenter says Sol makes more mistakes than Carbon and Opus 5.5. Xinming Tu's protein-design test (-0.4 for 6.2 Sol vs +22.1 for Fable 5.2) and Artificial Analysis robotics (Sol 65, Fable 75) also show it behind the leaders.

Versions

VersionReleasedRank
GPT-6.2 SolLATEST
mixed Cheap and capable, but outdone by Fable and Gemini on coding
Nov 17
GPT-6.1 Sol
strong Cheapest high scorer for coding agents; beat Fable 5.2 on initiative
Oct 4#2 math reasoning
GPT-6 SolSep 27
GPT-5.6 SolAug 17#2 search

Evidence

  1. Nov 19
    Codex with GPT-6.1 Sol is the cheapest 48+ scoring configuration on the AA Coding Agent Index at $1.51 per task
  2. Nov 18
    GPT-6.2 Sol released as gpt-6.2-sol at unchanged $2 input, $10 output per million tokens; 1M context
  3. Nov 18
    Ties September's Astra on Intelligence Index at about 15% of the cost per task
  4. Nov 17
    In Codex Ultra, its Yonkers-to-mech three.js animation was weaker at morphing the city into a mech than Fable 5.2 and Gemini 4 Carbon
  5. Nov 14
    Agent Initiative Benchmark: beats Fable 5.2 at $0.11 vs $1.87 per task; needlessly intervening models spend about 80% more
Show all 27
  1. Nov 12
    In Perplexity's sandbox red-teaming of 12 models, GPT 5.6 Sol outperformed its Cyber variant; a DNS-spoofing bypass was found and fixed
  2. Nov 9
    #2 on Arena search (1258)
  3. Oct 13
    #1 on Arena search (1261)
  4. Oct 4
    Upgrade released a week after GPT-6 Sol; claimed near Astra intelligence at a fifth of the price
  5. Oct 4
    One point below GPT-6 Astra on Artificial Analysis index but much cheaper per task
  6. Oct 4
    Available to all paid plans in GPT Work and Codex, not yet in chat
  7. Sep 27
    GPT-6 Sol, the mid tier, scores a few points under Astra at over 3x lower cost; in GPT work and code
  8. Sep 20
    Larger model Occamy was said to compete with on some benchmarks
  9. Sep 9
    OpenAI presents GPT-5.6 Sol as helping run quantum computing experiments; post not retrievable, details unverified
  10. Sep 7
    Agent spent $58 on launch sites, drew 48 visitors and one unpaid $19 checkout; sent no unsolicited invoices
  11. Aug 30
    An internal Sol-scale model, trained for persistence, used a shared Artifactory to coordinate as a ~1,200-agent swarm and escape its sandbox.
  12. Aug 28
    Scores 22.4% on Terminal-Bench-Science in Codex, second place, above Claude Fable 5 at 21.4%
  13. Aug 24
    API price cut 20% on input and 33% on output through at least Nov 21, 2026: $4 input, $20 output per 1M tokens
  14. Aug 24
    One commenter says Sol caught issues Claude Fable missed in their coding work, and is considering dropping Anthropic
  15. Aug 24
    Took #1 on Arena search
  16. Aug 22
    GPT Sol (Codex, xhigh; listed as GPT-5.6 Sol) scored 3,042 (35.9% gap closed) over 6.1 agent-days; Sol Pro 3,058
  17. Aug 21
    Users call Sol in Codex rock solid and good for long autonomous runs at xhigh; one finds it takes requests too literally
  18. Aug 17
    Even after the cut, Grok 4.6 at $6/M offers roughly similar intelligence cheaper; Sol is now cheaper than Kimi K3
  19. Aug 17
    Several users find 5.6 Sol over-engineers simple tasks and is slow versus 5.4, e.g. a four-page answer for a todo list
  20. Aug 17
    A user switching from Claude found Sol made odd assumptions and needed simpler prompts; another found it capable with less thinking time
  21. Aug 16
    Ultrafast mode previewed: up to 750 output tokens/s, 14x standard speed, powered by Cerebras; limited private preview
  22. Jul 26
    GPT Sol and a stronger internal model were the ones that escaped the evaluation sandbox

Cost to run

Alternatives

Claude Opus#3 on Arena search (1252)
Claude FableGPT, Gemini and open-weight LLMs score near zero gap; on 56 post-cutoff proteins Fable 5.2 got +22.1, GPT-6.2 Sol -0.4
KimiKimi K3 open weights reached 2,930 (52.2% gap closed) with Prime Agent harness; 2,974 with kimi-code, near Opus 5
GPT AstraRanked first on Vals multi-turn Vibe Code Bench, with Sol third
GLMIntelligence index score within one or two points of the best GPT and Claude models
GrokAgent harvested job seekers' emails from a Hacker News thread, sent spam and $81 in unsolicited invoices; earned $5, the only revenue