CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
← Back to feed

OpenRouter and Vercel Halve GPT-5.6 Sol to $2.50/$15 — Matching Kimi K3 on Output Cost

OpenRouter announced on X on August 17 that GPT-5.6 Sol’s call cost is halved on its platform. Input drops from $5 to $2.50 per million tokens; output drops from $30 to $15. No expiry date was given for the OpenRouter discount.

Vercel announced a matching offer the same day. Its 50% reduction covers both standard and Fast modes of GPT-5.6 Sol and runs through September 18, 2026.

What Changes at $15 Output

OpenAI’s official API pricing for Sol has not moved. Luna and Terra received official cuts of 80% and 20% respectively on July 30; Sol stayed at $5/$30 on the OpenAI platform. The platform discounts from OpenRouter and Vercel change the effective price without a change to the underlying list rate.

At $15 per million output tokens, GPT-5.6 Sol now costs the same as Kimi K3 on throughput — Kimi K3 lists at $3 input / $15 output on OpenRouter. On input, Sol at $2.50 undercuts Kimi K3 by 50 cents per million tokens.

DeepSeek V4 Flash remains far below both at roughly $0.22 input and $0.66 output per million tokens off-peak. DeepSeek V4 Pro is similarly priced at substantially below $1 per million output tokens. For organisations willing to accept open-weight deployment overhead, the cost gap to Sol is still roughly 15x on output.

Cost Per Task Shift

LiveBench’s cost-per-successful-task metric previously put GPT-5.6 Sol at $0.515 per task at standard pricing. At 50% off, the effective rate drops to approximately $0.26 — below Kimi K3’s $0.348 per task and well below Claude Opus 5 Thinking Max Effort’s $0.699. Claude Fable 5 Max Effort leads LiveBench overall at 83.0 but costs $1.439 per task; Sol at halved rates is now roughly 5.5x cheaper while trailing by only two overall points on the same benchmark.

Benchmark Context

GPT-5.6 Sol’s published scores at the time of the discount:

  • SWE-bench Verified: 96.2% (vals.ai neutral bash-only harness)
  • Terminal-Bench 2.1: 88.8%
  • LiveBench overall: 81.0, second behind Claude Fable 5 Max Effort (83.0)
  • Agentic coding on LiveBench: 56.2%

Sol’s SWE-bench position is #2 globally, behind Claude Opus 5 at 97.0% and ahead of Grok 4.6 at 95.6%.

Platform Discount vs Official Cut

This is not an OpenAI price reduction. OpenRouter and Vercel are either absorbing part of the margin or operating under volume arrangements not reflected in the public rate card. Whether this becomes a permanent rate change or a time-limited promotion to grow Sol’s share of platform call volume is not disclosed.

OpenAI’s builder’s guide, published August 13, notes “we’ve since reduced prices further” on the GPT-5.6 family and links to a separate price-performance update — suggesting the official rate card may also move, independently of platform-level actions.

For buyers routing through OpenRouter or building on Vercel, the effective date is August 17. Buyers on the direct OpenAI API continue to pay $5/$30 unless an official cut follows.